WO2020065807A1 - Information processing device, processing device, information processing method, processing method, determination method, and program - Google Patents

Information processing device, processing device, information processing method, processing method, determination method, and program Download PDF

Info

Publication number
WO2020065807A1
WO2020065807A1 PCT/JP2018/035851 JP2018035851W WO2020065807A1 WO 2020065807 A1 WO2020065807 A1 WO 2020065807A1 JP 2018035851 W JP2018035851 W JP 2018035851W WO 2020065807 A1 WO2020065807 A1 WO 2020065807A1
Authority
WO
WIPO (PCT)
Prior art keywords
recommended
information
combination
sensor
prediction
Prior art date
Application number
PCT/JP2018/035851
Other languages
French (fr)
Japanese (ja)
Inventor
江藤 力
Original Assignee
日本電気株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 日本電気株式会社 filed Critical 日本電気株式会社
Priority to PCT/JP2018/035851 priority Critical patent/WO2020065807A1/en
Priority to JP2020547702A priority patent/JP7056747B2/en
Publication of WO2020065807A1 publication Critical patent/WO2020065807A1/en

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01DMEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
    • G01D21/00Measuring or testing not otherwise provided for
    • G01D21/02Measuring two or more variables by means not covered by a single other subclass
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N5/00Analysing materials by weighing, e.g. weighing small particles separated from a gas or liquid
    • G01N5/02Analysing materials by weighing, e.g. weighing small particles separated from a gas or liquid by absorbing or adsorbing components of a material and determining change of weight of the adsorbent, e.g. determining moisture content

Definitions

  • the present invention relates to an information processing device, a processing device, an information processing method, a processing method, a determining method, and a program.
  • Patent Document 1 discloses an odor sensor provided with a plurality of sensor elements. Specifically, it is disclosed that a plurality of sensor elements are provided with a substance adsorption film having different characteristics, and each sensor element can take a configuration that exhibits a specific reaction to a molecule to be acted on. I have.
  • Patent Document 1 does not disclose how to select a combination of sensor elements according to the purpose of detection or a preferable detection environment.
  • An object of the present invention is to provide a technique for deriving a suitable combination of sensors or a preferable detection environment for a desired purpose.
  • the information processing device of the present invention includes: Based on execution results of machine learning using a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data, based on a recommended condition of a detection environment of the sensor and a combination of one or more of the sensors And recommendation information generating means for generating recommendation information in which the information is associated with.
  • the first determination method of the present invention is as follows: A method for determining the sensor to be used based on the recommendation information generated by the information processing apparatus of the present invention and information indicating the detection environment.
  • the second determination method of the present invention includes: This is a method for determining the detection environment based on the recommendation information generated by the information processing apparatus of the present invention and information indicating the available sensors.
  • the first processing device of the present invention comprises: Based on information indicating the detection environment and the recommended information relating the recommended conditions of the sensor detection environment and the recommended combination information indicating the combination of one or more sensors, and outputting the combination,
  • the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data.
  • the second processing device of the present invention comprises: Based on information indicating the recommended conditions of the detection environment of the sensor and recommended combination information indicating a combination of one or more sensors, and information indicating the available sensors, the recommended conditions are output.
  • the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data.
  • the information processing method of the present invention includes: Based on execution results of machine learning using a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data, based on a recommended condition of a detection environment of the sensor and a combination of one or more of the sensors And generating recommended information in which recommended information is associated with.
  • the first processing method of the present invention is as follows. Based on information indicating the detection environment and the recommended information relating the recommended conditions of the sensor detection environment and the recommended combination information indicating the combination of one or more sensors, and outputting the combination,
  • the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data.
  • the second processing method of the present invention includes: Based on information indicating the recommended conditions of the detection environment of the sensor and recommended combination information indicating a combination of one or more sensors, and information indicating the available sensors, the recommended conditions are output.
  • the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data.
  • the first program of the present invention is: A computer is caused to execute each step of the information processing method of the present invention.
  • a second program according to the present invention includes: The computer is caused to execute each step of the processing method of the present invention.
  • FIG. 2 is a diagram illustrating a configuration of the information processing apparatus according to the first embodiment. It is a figure which illustrates a sensor. It is a figure which illustrates time series data. It is a figure which illustrates sensor output data from a set of a plurality of types of sensors.
  • 5 is a flowchart illustrating an information processing method according to the first embodiment.
  • FIG. 3 is a diagram illustrating a prediction model used for machine learning performed by a prediction formula generation unit according to the first embodiment.
  • FIG. 2 is a diagram illustrating a computer for realizing an information processing device.
  • FIG. 9 is a diagram illustrating a configuration of an information processing apparatus according to a second embodiment. 9 is a flowchart illustrating an information processing method according to a second embodiment.
  • FIG. 13 is a flowchart illustrating a processing method according to a third embodiment. It is a figure which illustrates the composition of the processing device concerning a 4th embodiment. 13 is a flowchart illustrating a processing method according to a fourth embodiment.
  • each component of each device is not a configuration of a hardware unit but a block of a functional unit, unless otherwise specified.
  • Each component of each device is composed mainly of a CPU of an arbitrary computer, a memory, a program for realizing the components of this drawing loaded in the memory, a storage medium such as a hard disk for storing the program, and a network connection interface. It is realized by any combination of software and software. There are various modifications in the method and apparatus for realizing the method.
  • FIG. 1 is a diagram illustrating a configuration of an information processing device 20 according to the first embodiment.
  • the information processing device 20 according to the present embodiment includes the recommended information generation unit 270.
  • the recommendation information generating means 270 generates recommendation information based on a result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data.
  • the recommended information is information that associates a recommended condition of a sensor detection environment with a combination of one or more sensors. This will be described in detail below.
  • the information processing apparatus 20 further includes a prediction formula generation unit 210 and an extraction unit 220.
  • the prediction formula generation means 210 generates a prediction formula for predicting an odor component, which is a formula using a plurality of feature amounts as variables by performing machine learning.
  • the extracting unit 220 extracts one or more sensors from the set based on a plurality of weights for a plurality of feature amounts in the prediction formula. More specifically, the extraction unit 220 extracts a sensor that is an output source of a feature amount that is weighted with a weight that satisfies or does not satisfy a predetermined condition among a plurality of weights in the prediction formula. Then, the extracting unit 220 generates recommended combination information indicating the combination of the extracted sensors.
  • the recommended information includes one or more recommended combination information.
  • FIG. 2 is a diagram illustrating the sensor 10.
  • the sensor 10 has a receptor to which a molecule is attached, and a detection value changes according to attachment and detachment of the molecule at the receptor.
  • the gas sensed by the sensor 10 is referred to as a target gas.
  • the time-series data of the detection values output from the sensor 10 is referred to as time-series data 14.
  • the time-series data 14 is also described as Y
  • the detected value at the time t is also described as y (t).
  • Y is a vector in which y (t) is enumerated.
  • the senor 10 is a membrane-type surface stress sensor (MSS).
  • MSS membrane-type surface stress sensor
  • the MSS has, as a receptor, a functional film to which a molecule is attached, and the stress generated in a support member of the functional film is changed by attachment and detachment of the molecule to and from the functional film.
  • the MSS outputs a detection value based on the change in the stress.
  • Various materials such as organic, inorganic, and bio-based materials can be used for the MSS functional film.
  • the target molecule to which the sensor 10 responds and the response characteristics depend on the functional film. Therefore, by combining a plurality of types of sensors 10 having different functional films from each other, it becomes possible to analyze a complicated odor composed of a mixed gas containing various components.
  • the sensor 10 is not limited to the MSS, and changes in physical quantities related to the viscoelasticity and dynamic characteristics (mass, moment of inertia, etc.) of the members of the sensor 10 that occur in response to attachment and detachment of molecules to and from the receptor. Any type of sensor may be used as long as it outputs a detection value based on the above, and various types of sensors such as a cantilever type, a film type, an optical type, a piezo, and a vibration response can be adopted. Also in these sensors 10, a plurality of types of sensors 10 having different target molecules to which the sensor 10 responds and at least one of the response characteristics can be combined. Note that the plurality of types of sensors 10 may detect information having the same attribute (such as the mass of an attached molecule).
  • the types of the sensors 10 are many.
  • the number of sensors 10 that can be actually used in the detection device is limited. Therefore, it is necessary to select which type of sensor 10 should be used in combination to perform the target detection.
  • the output of the sensor 10 also depends on its detection environment. Thus, depending on the environment, preferred sensor 10 combinations may be different. Similarly, depending on the combination of sensors 10 used, the conditions of the preferred detection environment may vary.
  • the recommended information in which the recommended detection environment and the combination of the sensor 10 are associated with each other can be obtained. Then, based on the recommended information, a preferable combination of the sensors 10 and detection in a detection environment can be performed.
  • the prediction formula generation unit 210 performs prediction regarding an odor component by performing machine learning using a plurality of feature amounts based on outputs from a set of a plurality of types of sensors 10 and correct data as inputs. Generate a prediction equation for.
  • the prediction expression is an expression using a plurality of feature amounts as variables, and the weight for each feature amount in the prediction expression corresponds to the magnitude of the contribution of the feature amount to the prediction result. Therefore, the extracting unit 220 can determine the sensor 10 that has a large contribution to the purpose and the sensor 10 that has a small contribution based on the information indicating the prediction formula.
  • the prediction formula generation means 210 generates a prediction formula using a model including a branch based on the detection environment. By doing so, a preferable combination of the sensors 10 is derived in association with the detection environment.
  • the prediction formula generation unit 210 associates the recommended combination information with a detection environment condition suitable for the prediction formula, which is a detection environment condition based on a branch condition, as a recommended condition. Then, recommended information including a plurality of sets of the recommended combination information and the recommended condition associated with each other is generated.
  • the feature amount and the prediction formula will be described in detail below.
  • the feature amount is a value obtained based on the output of the sensor 10. However, one or more feature amounts are obtained for one sensor 10, and each feature amount depends only on the output of one sensor 10.
  • the time-series data 14 is time-series data in which the detection values output by the sensor 10 are arranged in ascending order of the time output from the sensor 10.
  • the time-series data 14 may be obtained by subjecting the time-series data of the detection value obtained from the sensor 10 to predetermined preprocessing.
  • pre-processing for example, filtering for removing a noise component from the time-series data can be employed.
  • FIG. 3 is a diagram illustrating the time-series data 14.
  • the time-series data 14 is obtained by exposing the sensor 10 to a target gas.
  • the time-series data 14 may be obtained by an operation of exposing the sensor 10 to the gas to be measured and an operation of removing the gas to be measured from the sensor 10.
  • data of the period P1 is obtained by exposing the sensor 10 to the target gas
  • data of the period P2 is obtained by an operation of removing the gas to be measured from the sensor 10.
  • the operation of removing the gas to be measured from the sensor 10 includes, for example, an operation of exposing the sensor 10 to a purge gas.
  • the operation of exposing the sensor 10 to the gas to be measured and the operation of removing the gas to be measured from the sensor 10 may be repeated to obtain a plurality of time-series data 14.
  • FIG. 4 is a diagram illustrating sensor output data 16 from a set 100 of a plurality of types of sensors 10.
  • a set 100 of the sensors 10 includes a first sensor 10a, a second sensor 10b, a third sensor 10c, and a fourth sensor 10d.
  • the set 100 is modularized, and measurement is performed on the same target gas in the same detection environment.
  • the set 100 of sensors 10 comprises a plurality of sensors 10 arbitrarily selected from a large number of available sensors 10.
  • the sensor output data 16 is data obtained by combining the time-series data 14 obtained from each of the plurality of types of sensors 10.
  • the sensor output data 16 is obtained by sequentially arranging the time series data 14 of the first sensor 10a, the second sensor 10b, the third sensor 10c, and the fourth sensor 10d.
  • the feature amount vector X is a vector having a plurality of feature amounts as elements.
  • xj may be a numerical value or a vector. If x j is a vector, x j is a vector whose elements a plurality of feature quantity based on the output of the same sensor 10.
  • the feature amount xj is, for example, the time-series data 14 of the sensor 10, data obtained by differentiating the time-series data 14, or a set ⁇ ⁇ of contribution values described later.
  • the prediction formula generation unit 210 can acquire the time-series data 14 or the sensor output data 16 and calculate a feature amount based on the acquired data. However, instead of acquiring the time-series data 14 or the sensor output data 16, the prediction formula generation unit 210 may acquire a feature amount derived outside the information processing device 20.
  • W is a vector and b is a constant.
  • Each element of the weight W is a coefficient for each element of the feature amount vector X.
  • the obtained z indicates the prediction result.
  • the prediction formula may be used for discrimination or may be used for regression prediction. For example, in the prediction formula used to determine the presence or absence of a certain odor component, if z is equal to or greater than a predetermined criterion, it is determined that the gas to be measured contains the odor component to be detected and is smaller than the criterion. In this case, it can be determined that the gas to be measured does not contain the odor component to be detected.
  • the regression prediction include prediction of manufacturing quality based on the smell of a product such as a beverage, and prediction of a body state by measuring breath.
  • time-series data 14 the sensor output data 16, the feature amounts, and the prediction formulas are examples, and the time-series data 14, the sensor output data 16, the feature amounts, and the forms of the prediction formulas according to the present embodiment. Is not limited to the above.
  • a set ⁇ ⁇ of contribution values which is an example of a feature value, will be described below.
  • the sensing by the sensor 10 is modeled as follows. (1) The sensor 10 is exposed to a target gas containing K kinds of molecules. (2) The concentration of each molecule k in the target gas is constant ⁇ k . (3) The sensor 10 can adsorb a total of N molecules. (4) The number of molecules k attached to the sensor 10 at time t is n k (t).
  • the change over time of the number n k (t) of molecules k attached to the sensor 10 can be formulated as follows.
  • the first and second terms on the right side of the equation (1) are the increase amount (the number of molecules k newly attached to the sensor 10) and the decrease amount (the molecule k detached from the sensor 10) per unit time. Number). Further, ⁇ k and ⁇ k are a rate constant representing the rate at which the molecule k adheres to the sensor 10 and a rate constant representing the rate at which the molecule k separates from the sensor 10, respectively.
  • the concentration ⁇ k is constant
  • the number n k (t) of the numerator k at the time t can be formulated from the above equation (1) as follows.
  • n k (t) is expressed as follows.
  • the detection value of the sensor 10 is determined by the stress applied to the sensor 10 by molecules contained in the target gas. Then, it is considered that the stress acting on the sensor 10 by a plurality of molecules can be represented by a linear sum of the stress acting on each molecule. However, it is considered that the stress generated by the molecule differs depending on the type of the molecule. That is, it can be said that the contribution of the molecule to the detection value of the sensor 10 differs depending on the type of the molecule.
  • the detection value y (t) of the sensor 10 can be formulated as follows.
  • both ⁇ k and k k represent the contribution of the numerator k to the detection value of the sensor 10. Note that “rising” corresponds to the above-described period P1, and “falling” corresponds to the above-described period P2.
  • the time-series data 14 obtained from the sensor 10 sensing the target gas can be decomposed as in the above equation (4), the types of molecules contained in the target gas and each type of molecule are contained in the target gas.
  • the set of feature constants ⁇ may be determined in advance or may be generated by the information processing device 20.
  • i i is a contribution value representing the contribution of the characteristic constant ⁇ i to the detection value of the sensor 10.
  • a contribution value ⁇ i representing the contribution of each feature constant ⁇ i to the time series data 14 is calculated.
  • the ⁇ set of contribution value xi] i can be a feature quantity representing the feature of the target gas.
  • the feature quantity of the target gas does not necessarily have to be represented as a vector.
  • equation (5) can be expressed as follows.
  • the contribution of the molecule to the detection value of the sensor 10 is considered to be different depending on the type of the molecule. Therefore, the set ⁇ ⁇ of the contribution values described above depends on the type of the molecule contained in the target gas and the mixing ratio thereof. Are likely to be different. Therefore, the set ⁇ ⁇ of contribution values can be used as information that can distinguish a gas in which a plurality of types of molecules are mixed, that is, as a feature amount of the gas.
  • Using the ⁇ set of contribution values ⁇ as the feature of the target gas has other advantages besides the advantage of being able to handle gases containing multiple types of molecules.
  • the degree of similarity between gases can be easily grasped. For example, if the feature amount of the target gas is represented by a vector, the degree of similarity between the gases can be easily grasped based on the distance between the feature vectors.
  • Using the ⁇ set of contribution values ⁇ as the feature quantity has the advantage that it is possible to make the time constant change and the change in the mixture ratio robust against the change in the mixture ratio.
  • the “robustness” here is a property that “when the measurement environment or the measurement target slightly changes, the obtained feature amount also slightly changes”.
  • the characteristic amount will also gradually change. This property can be seen from the fact that in equation (4), the contribution value ⁇ k is proportional to ⁇ k representing the gas concentration, so that a small change in the concentration appears as a small change in the contribution value.
  • FIG. 5 is a flowchart illustrating an information processing method according to the first embodiment.
  • the information processing method according to the present embodiment includes a recommended information generating step S270.
  • recommended information is generated based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from the set 100 of a plurality of types of sensors 10 and correct data.
  • the information processing method further includes a prediction formula generation step S210 and an extraction step S220.
  • a prediction formula is generated by performing machine learning.
  • the extraction step S220 one or more sensors 10 are extracted from the set 100 based on a plurality of weights for a plurality of feature amounts in the prediction formula. This will be described in detail below.
  • the plurality of feature amounts to be input for machine learning are obtained from the result of measuring a known target gas with the set 100 of the sensors 10 and are, for example, the feature amount vector X described above.
  • the prediction formula generation means 210 acquires the time series data 14, the sensor output data 16, or the feature quantity vector X.
  • the prediction formula generation means 210 may acquire the time-series data 14, the sensor output data 16, or the feature vector X from a storage device accessible from the prediction formula generation means 210, or a device external to the information processing apparatus 20. From the sensor 10 or from the sensor 10.
  • the prediction formula generation unit 210 calculates the feature quantity vector X based on these data.
  • the feature quantity vector X may be obtained by in-situ measurement, or may be prepared in advance and stored in a storage device.
  • the prediction formula generation unit 210 acquires correct data for the feature vector X.
  • the correct answer data is information indicating a prediction result to be obtained by a prediction formula for the associated feature amount vector X. That is, the correct answer data is information corresponding to the measured known target gas.
  • the correct answer data may be input to the information processing device 20 by the user, or may be stored in advance in the storage device accessible from the prediction formula generation unit 210 in association with the feature amount vector X (that is, a plurality of feature amounts). good.
  • the prediction formula generation means 210 acquires a detection environment associated with the feature amount.
  • This detection environment is a detection environment when the time-series data 14 that is the basis of the feature amount is obtained.
  • the detection environment is not particularly limited, for example, temperature, humidity, atmospheric pressure, type of impurity gas, type of purge gas, sampling period of odor component, distance between target and sensor 10, at least one of objects existing around sensor 10 Including any.
  • the temperature, the humidity, and the atmospheric pressure are the temperature, the humidity, and the atmospheric pressure around the sensor 10, and more specifically, the temperature, the humidity, and the atmospheric pressure of the atmosphere surrounding the functional film of the sensor 10.
  • the type of the contaminant gas is the type of gas supplied to the sensor 10 together with the target odor component in the operation of exposing the sensor 10 to the target gas.
  • examples of the type of impurity gas include an inert gas such as nitrogen, and air.
  • the type of the purge gas is a gas supplied to the sensor 10 in the operation of removing the gas to be measured from the sensor 10.
  • examples of the purge gas include an inert gas such as nitrogen, and air.
  • the sampling cycle of the odor component is a repetition cycle when the operation of exposing the sensor 10 to the gas to be measured and the operation of removing the gas to be measured from the sensor 10 are repeatedly performed.
  • the distance between the target object and the sensor 10 is the distance between the target object and the sensor 10 when the sensor 10 is placed around a specific target object to perform detection.
  • the object existing around the sensor 10 is the type of the target when the sensor 10 is placed around a specific target to perform detection.
  • the prediction formula generation unit 210 performs machine learning using a learning data set including a plurality of feature amounts, correct answer data, and a detection environment, which are associated with each other, as inputs.
  • the prediction formula generation unit 210 can improve the accuracy of the prediction formula by performing machine learning using a plurality of learning data sets.
  • the plurality of learning data sets as described above, in the measurement of the target gas by the sensor 10, the operation of exposing the sensor 10 to the gas to be measured and the operation of removing the gas to be measured from the sensor 10 are repeatedly performed. Obtained by:
  • the prediction formula generation means 210 ends the learning when, for example, a predetermined number of learning iterations (the number of learning data sets) is satisfied.
  • the plurality of learning data sets include two or more learning data sets whose detection environments are different from each other. By doing so, the conditions for branching are appropriately derived by heterogeneous learning, and a model is generated.
  • the feature amount used for machine learning may be obtained by simulating the response of the sensor 10 to the target gas.
  • a plurality of learning data sets can be generated using results obtained under simulation conditions with different detection environments. However, when a plurality of different simulation results are obtained for the same detection environment, a plurality of learning data sets may be generated including the results obtained under the same simulation conditions.
  • FIG. 6 is a diagram illustrating a prediction model used for machine learning performed by the prediction formula generation unit 210 according to the present embodiment.
  • the machine learning is a heterogeneous mixture learning in which, in addition to a plurality of feature amounts and correct answer data, a detection environment associated with the feature amounts is further input.
  • the branch condition in the model is generated by heterogeneous learning.
  • Models used for machine learning have a hierarchical structure that includes a plurality of nodes.
  • a branch formula is located at one or more intermediate nodes as a branch condition, and a prediction formula is located at the lowest anode.
  • condition A, condition B1 and condition B2 are branch conditions, and equations 1 to 4 are prediction equations, respectively.
  • the specific configuration of the model such as the number of intermediate nodes and the number of anodes is not particularly limited.
  • the prediction formula generation means 210 generates a specific model including one or more prediction formulas and branch conditions by performing machine learning in the prediction formula generation step S210. Specifically, the prediction formula generation unit 210 derives a weight W and a constant b as information indicating the prediction formula. Further, the prediction formula generation unit 210 derives information indicating the configuration of the model and the conditions of each branch included in the model.
  • the condition of the detection environment which is a premise is linked to each prediction formula as a recommended condition.
  • Each prediction equation is particularly effective in an environment that satisfies the recommended conditions associated with the prediction equation.
  • the recommended condition is based on a branch condition in a model generated simultaneously with the prediction formula. More specifically, the recommended conditions are determined in the generated model by the conditions of the branches that pass from the start to the prediction formula of the anode and the determination results. For example, in the example of this drawing, when the condition A is “temperature> T 1 ” and the condition B 2 is “humidity> H 1 ”, the recommended condition associated with the equation 3 is “the temperature is T 1 or less, and humidity is higher than H 1 ".
  • a specific model including a branch condition used in machine learning may be set by a user instead of being generated by machine learning.
  • the machine learning need not be heterogeneous learning.
  • a branching condition can be repeatedly updated together with a prediction formula during repetition of learning, but a model obtained at a stage during learning may be fixed and used in subsequent learning.
  • the extraction step 220 is performed by the extraction means 220.
  • the extraction unit 220 extracts the sensor 10 having a high contribution to the prediction result in the prediction formula based on the weight in each prediction formula and a predetermined condition regarding the weight. Specifically, the extraction unit 220 acquires information indicating the prediction formula from the prediction formula generation unit 210. Then, the magnitude of the weight for the characteristic amount of each sensor 10 indicated in the information indicating the prediction formula is calculated.
  • w j may be a numerical value or a vector.
  • each element of w j is a weight for each feature amount that is an element of x j .
  • the magnitude of the weight is, for example, the norm of w j .
  • the magnitude of the weight is the absolute value of w j .
  • the extraction means 220 further determines whether or not the magnitude of the calculated weight satisfies a predetermined condition.
  • the information indicating the condition is stored in a storage device accessible from the extracting unit 220 in advance. For example, when the condition indicates a condition for the sensor 10 having a high degree of contribution to the prediction result, such as “the magnitude of the weight is equal to or greater than the reference value”, the extracting unit 220 determines the sensor 10 corresponding to the weight satisfying this condition. Extract. On the other hand, if the condition indicates a condition of the sensor 10 that makes a small contribution to the prediction result, such as “the magnitude of the weight is equal to or less than the reference value”, the extracting unit 220 extracts the sensor 10 corresponding to the weight that does not satisfy the condition. I do.
  • the number of sensors 10 to be extracted is not particularly limited. Then, the extracting unit 220 generates recommended combination information indicating the combination of the extracted sensors 10. The generated recommended combination information is associated with information indicating a prediction formula.
  • the extraction unit 220 similarly generates combination information for all the prediction expressions generated by the prediction expression generation unit 210.
  • the extracting means 220 may extract the sensor 10 based on the weight for the feature based on only part of the data in the period P1 and the period P2. Specifically, in each of the period P1 and the period P2, the sensor 10 may be extracted based on the weight for the feature amount based on the data from the beginning of the period to after a predetermined time.
  • the recommended information generating means 270 performs a recommended information generating step S270.
  • the recommended information generating unit 270 acquires, from the extracting unit 220, information indicating a prediction formula, recommended conditions, and recommended combination information, which are associated with each other.
  • the recommended information generating means 270 generates recommended information including the acquired information, for example. That is, the recommended information includes the recommended combination information, the information indicating the prediction formula, and the recommended condition in a state where they are associated with each other. Note that the recommended information only needs to include at least the recommended combination information and the recommended conditions in a state where they are associated with each other.
  • the number of pieces of recommended combination information acquired by the recommended information generation unit 270 and included in the recommended information depends on the number of prediction formulas generated by the prediction formula generation unit 210. . That is, it depends on the configuration of the model generated by the prediction expression generation means 210.
  • the recommended information generating unit 270 acquires all the recommended combination information generated by the extracting unit 220, and all the recommended combination information is associated with information indicating a prediction formula and a recommended condition.
  • the generated recommended information may be stored in a storage device accessible from the recommended information generating unit 270, may be output to an external device, or may be presented to the user on a display device or the like.
  • the recommendation information generation unit 270 adds the new recommendation combination information acquired from the extraction unit 220 to the existing recommendation information held in the storage device accessible from the recommendation information generation unit 270 in the recommendation information generation step S270. New recommended information is generated by adding recommended conditions and the like. Then, the recommended information of the storage device is updated.
  • the senor 10 to be used can be determined based on the recommended information generated by the information processing device 20 and the information indicating the detection environment. For example, when a user intends to produce a sensor module used in a specific detection environment, it is necessary to select the sensor 10 from a large number of sensors 10 within a range that can be mounted on the sensor module. Therefore, the user finds a recommended condition corresponding to the target detection environment from the recommended information. Then, a combination of the sensors 10 mounted on the sensor module is determined based on the recommended combination information associated with the recommended condition.
  • the detection environment can be determined based on the recommended information generated by the information processing device 20 and the information indicating the available sensors 10. For example, when the user can use one or more specific sensors 10, the recommended information is used to determine under what detection environment the measurement should be performed in order to perform a target prediction using the sensors 10. You can know. Specifically, the user finds recommended combination information that can be realized by a combination of available sensors 10. Then, the user understands that it is preferable to perform the measurement within the range of the recommended condition associated with the recommended combination information.
  • the user can make a prediction regarding the odor component in accordance with the finally adopted recommended combination information and the information indicating the prediction formula associated with the recommended condition. Specifically, in the prediction regarding the odor component, a feature amount is calculated based on the outputs from the plurality of sensors 10, and the feature amount is applied to the prediction formula. Then, a prediction result is obtained based on the value calculated by the prediction formula.
  • the information processing device 20 may generate a plurality of pieces of recommended information for each purpose of prediction.
  • Each functional component of the information processing device 20 may be implemented by hardware (eg, a hard-wired electronic circuit or the like) that implements each functional component, or a combination of hardware and software (eg: Electronic circuit and a program for controlling the same).
  • hardware eg, a hard-wired electronic circuit or the like
  • software eg: Electronic circuit and a program for controlling the same.
  • FIG. 7 is a diagram illustrating a computer 1000 for realizing the information processing device 20.
  • the computer 1000 is an arbitrary computer.
  • the computer 1000 is a stationary computer such as a personal computer (PC) or a server machine.
  • the computer 1000 is a portable computer such as a smartphone or a tablet terminal.
  • the computer 1000 may be a dedicated computer designed to realize the information processing device 20, or may be a general-purpose computer.
  • the computer 1000 has a bus 1020, a processor 1040, a memory 1060, a storage device 1080, an input / output interface 1100, and a network interface 1120.
  • the bus 1020 is a data transmission path through which the processor 1040, the memory 1060, the storage device 1080, the input / output interface 1100, and the network interface 1120 mutually transmit and receive data.
  • a method for connecting the processors 1040 and the like to each other is not limited to a bus connection.
  • the processor 1040 is various processors such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), and an FPGA (Field-Programmable Gate Array).
  • the memory 1060 is a main storage device realized using a RAM (Random Access Memory) or the like.
  • the storage device 1080 is an auxiliary storage device realized using a hard disk, an SSD (Solid State Drive), a memory card, or a ROM (Read Only Memory).
  • the input / output interface 1100 is an interface for connecting the computer 1000 and an input / output device.
  • an input device such as a keyboard and an output device such as a display device are connected to the input / output interface 1100.
  • the sensor 10 is connected to the input / output interface 1100.
  • the sensor 10 does not necessarily need to be directly connected to the computer 1000.
  • the sensor 10 may store the time-series data 14 in a storage device shared with the computer 1000.
  • the network interface 1120 is an interface for connecting the computer 1000 to a communication network.
  • the communication network is, for example, a LAN (Local Area Network) or a WAN (Wide Area Network).
  • the method by which the network interface 1120 connects to the communication network may be a wireless connection or a wired connection.
  • the storage device 1080 stores a program module that implements each functional component of the information processing apparatus 20.
  • the processor 1040 realizes a function corresponding to each program module by reading out each of these program modules into the memory 1060 and executing them.
  • the information processing apparatus 20 it is possible to generate recommendation information indicating a preferable combination of the combination of the sensors 10 and the conditions of the detection environment. Therefore, it is possible to know the combination of the sensors 10 and the environment appropriate for the purpose.
  • FIG. 8 is a diagram illustrating a configuration of an information processing device 20 according to the second embodiment.
  • FIG. 9 is a flowchart illustrating an information processing method according to the second embodiment.
  • the information processing device 20 according to the present embodiment is the same as the information processing device 20 according to the first embodiment except for the points described below.
  • the information processing apparatus 20 further includes a prediction accuracy calculation unit 230 that calculates the prediction accuracy of the prediction formula, and an evaluation unit 240 that evaluates the combination of the sensors 10.
  • the information processing method further includes a prediction accuracy calculation step S230 and an evaluation step S240.
  • the information processing apparatus 20 according to the present embodiment may not include at least one of the prediction accuracy calculation unit 230 and the evaluation unit 240. Further, the information processing method according to the present embodiment may not include at least one of the prediction accuracy calculation step S230 and the evaluation step S240.
  • the processing of the prediction accuracy calculation step S230 is performed by the prediction accuracy calculation means 230.
  • the timing at which the processing of the prediction accuracy calculation step S230 is performed is not particularly limited as long as it is after the prediction expression generation step S210 and before the evaluation step S240 described later.
  • the timing at which the processing of the prediction accuracy calculation step S230 is performed may be after the prediction expression generation step S210 and before the recommendation information generation step S270.
  • the prediction accuracy calculation means 230 calculates the prediction accuracy of each prediction formula generated by the prediction formula generation means 210.
  • a data set similar to the learning data set is used as the evaluation data set. That is, the evaluation data set includes a plurality of feature amounts, correct answer data, and a detection environment.
  • the plurality of learning data sets and the plurality of evaluation data sets do not include exactly the same data sets.
  • a part of a plurality of different data sets generated outside or inside the information processing device 20 may be used as a plurality of learning data sets, and the rest may be used as a plurality of evaluation data sets.
  • Prediction accuracy is the regression accuracy for prediction based on regression, for example, least squares error or mean squared error (RMSE).
  • the prediction accuracy is the determination accuracy for prediction based on the determination, and is, for example, a precision, a recall, an F value, a correct answer rate, or ROC_AUC.
  • the prediction accuracy calculation unit 230 can obtain or generate a plurality of evaluation data sets in the same manner as the prediction expression generation unit 210 obtains or generates a learning data set.
  • the prediction accuracy calculation means 230 obtains a prediction result by inputting the feature amount included in the evaluation data set into a prediction expression whose accuracy is to be evaluated. Then, it is determined whether or not the obtained prediction result matches the correct answer data included in the evaluation data set. Then, the prediction accuracy calculation means 230 performs the same processing for a plurality of evaluation data sets, and calculates the probability that the prediction result matches the correct answer data as the prediction accuracy of the prediction formula.
  • the calculated prediction accuracy is associated with the prediction formula.
  • the plurality of evaluation data sets may be based on measurement results in mutually different detection environments. However, for each prediction formula, only the evaluation data set obtained in an environment that satisfies the condition of the detection environment associated with the prediction formula is used for calculating the prediction accuracy.
  • the evaluation means 240 evaluates the combination of the sensors 10 based on, for example, at least one of the prediction accuracy of a prediction formula used when adopting the combination and the detection environment and the cost when adopting the combination. In particular, it is preferable that the evaluation unit 240 evaluates the combination of the sensors 10 based at least on the cost when the combination of the sensors 10 indicated in the recommended combination information is adopted.
  • Cost includes initial cost and running cost, for example.
  • Examples of the initial cost include a manufacturing cost and a procurement cost of the sensor 10.
  • the running costs include management costs, replacement costs caused by deterioration of the sensor 10, and human labor in handling.
  • a parameter indicating the cost of each sensor 10 is stored in advance in a storage device accessible by the evaluation unit 240, and the evaluation unit 240 acquires a parameter indicating the cost of the sensor 10 included in the combination from the storage device. Then, the parameters indicating the costs for all the sensors 10 included in the combination are added up to obtain a total value.
  • the evaluation unit 240 acquires the prediction accuracy of the prediction formula associated with the recommended combination information from the prediction accuracy calculation unit 230.
  • Evaluation means 240 further evaluates the combination using an evaluation function.
  • the evaluation function is a function that calculates an evaluation value based on one or more factors. Specifically, the evaluation function is represented by a linear sum of evaluation parameters indicating the evaluation result of each factor. For example, the evaluation parameter with the cost as the factor is the total value calculated as described above, and the evaluation parameter with the accuracy as the factor is the prediction accuracy acquired from the prediction accuracy calculation unit 230.
  • each evaluation parameter is multiplied by a coefficient to balance the weight for each factor with respect to the evaluation result or to determine the directionality of the evaluation. The coefficient is determined for each type of evaluation parameter.
  • Evaluation means 240 calculates an evaluation value as an evaluation result, for example, by applying the sum of parameters indicating cost and prediction accuracy to the evaluation function.
  • the evaluation result obtained by the evaluation means 240 increases as the sum of the costs decreases, and increases as the prediction accuracy improves.
  • Information indicating the evaluation function is stored in a storage device accessible by the evaluation unit 240 in advance. The calculated evaluation value is associated with the recommended combination information.
  • the evaluation means 240 may further evaluate the combination of the sensors 10 based on the number of the sensors 10 included in the combination. For example, when the number of sensors 10 included in the combination is a factor, for example, the number of sensors 10 can be an evaluation parameter in an evaluation function. Note that the evaluation result obtained by the evaluation unit 240 increases as the number of sensors 10 included in the combination decreases.
  • ⁇ Evaluation means 240 may evaluate a combination further based on recommendation conditions linked with recommended combination information. For example, when the size of the recommended condition is a factor, for example, the width of the range of the temperature, humidity, atmospheric pressure, cycle, distance, etc. indicated as the recommended condition, or the number of options of gas or object is evaluated by the evaluation function. Can be a parameter. When the practicality of the recommended condition is a factor, the distance between the center value of the range of temperature, humidity, atmospheric pressure, cycle, distance, etc. indicated as the recommended condition and a predetermined standard value is evaluated by the evaluation function. Can be a parameter. That is, it can be said that the shorter the distance, the higher the practicality. Note that the evaluation result obtained by the evaluation means 240 increases as the recommended conditions are wider, and increases as the practicality of the recommended conditions increases.
  • the recommended information generating means 270 performs a recommended information generating step S270.
  • the recommended information generating means 270 acquires information indicating a prediction formula, recommended conditions, recommended combination information, and an evaluation result, which are associated with each other.
  • the recommended information generating unit 270 generates recommended information including the acquired information. That is, the recommended information includes the recommended combination information, the information indicating the prediction formula, the recommended condition, and the evaluation result in a state where they are associated with each other.
  • the recommended information further includes the evaluation result of the evaluation unit 240 associated with the recommended combination information. However, the recommended information is replaced with the evaluation result or in addition to the evaluation result. And the prediction accuracy of the prediction formula associated with the recommended combination information.
  • the recommended information generating means 270 may select a prediction formula based on at least one of the prediction accuracy and the evaluation result, and only the information associated with the selected prediction formula may be included in the recommended information. Specifically, the recommendation information generation unit 270 selects a prediction expression having a higher prediction accuracy than a predetermined reference from the prediction expressions generated by the prediction expression generation unit 210. Alternatively, the recommendation information generation unit 270 selects a prediction expression of an evaluation result superior to a predetermined criterion from among the prediction expressions generated by the prediction expression generation unit 210. Then, the recommended information generating unit 270 generates recommended information including information (recommended combination information, recommended conditions, and the like) associated with the selected prediction formula.
  • the information processing apparatus 20 can also be realized by the computer 1000 as shown in FIG.
  • the storage device 1080 further stores program modules for realizing the prediction accuracy calculation unit 230 and the evaluation unit 240 of the information processing device 20.
  • the prediction accuracy of the prediction formula is calculated by the prediction accuracy calculation unit 230 or the evaluation by the evaluation unit 240 is performed, so that the usefulness of a plurality of pieces of recommended combination information can be compared with each other.
  • FIG. 10 is a diagram illustrating a configuration of a processing device 30 according to the third embodiment.
  • the processing device 30 according to the present embodiment, based on recommended information in which the recommended conditions of the detection environment of the sensor 10 and the recommended combination information indicating a combination of one or more sensors 10 are associated, and information indicating the detection environment, Output the combination.
  • the recommendation information is information based on an execution result of machine learning that inputs a plurality of feature amounts based on outputs from a set 100 of a plurality of types of sensors 10 and correct answer data.
  • the processing device 30 includes an extraction unit 320, a selection unit 340, and an output unit 370.
  • the extracting unit 320 extracts, from the recommended conditions included in the recommended information, a recommended condition to which the information indicating the detection environment matches.
  • the selection unit 340 selects one or more pieces of recommended combination information from the recommended combination information associated with the extracted recommended conditions. Then, the output unit 370 outputs the combination indicated by the selected recommended combination information.
  • FIG. 11 is a flowchart illustrating a processing method according to the third embodiment.
  • the combination is output based on the recommended information in which the recommended conditions of the sensor detection environment and the recommended combination information indicating the combination of one or more sensors 10 are associated with each other and the information indicating the detection environment.
  • the processing method includes an extraction step S320, a selection step S340, and an output step S370.
  • the extraction step S320 a recommended condition to which the information indicating the detection environment matches is extracted from the recommended conditions included in the recommended information.
  • the selection step S340 one or more recommended combination information is selected from the recommended combination information associated with the extracted recommended condition.
  • the output step S370 the combination indicated by the selected recommended combination information is output.
  • the processing method according to the present embodiment is realized by the processing device 30.
  • the recommended information according to the present embodiment is the same as the recommended information generated by the recommended information generating unit 270 according to at least one of the first and second embodiments, for example. Further, in the following description, the information processing device 20 is the same as the information processing device 20 according to at least one of the first and second embodiments. According to the processing device 30 and the processing method according to the present embodiment, a preferable combination of the sensors 10 can be obtained from the detection environment in which the sensor 10 is to be used, using the recommended information. This will be described in detail below.
  • the extraction means 320 acquires recommended information and information indicating the detection environment.
  • the extracting unit 320 can obtain the recommended information from a storage device or the information processing device 20 accessible from the extracting unit 320. Further, the user of the processing device 30 can input information indicating the detection environment to the processing device 30. Then, the extracting unit 320 acquires information indicating the input detection environment.
  • the information indicating the detection environment indicates, for example, the detection environment when the user intends to perform measurement using any of the sensors 10.
  • the recommended information includes one or more recommended combination information and a recommended condition associated therewith.
  • the extracting unit 320 extracts, from the recommended information, all the recommended conditions to which the information indicating the detection environment matches.
  • the information indicating the recommended condition and the detection environment may include a plurality of elements such as temperature and humidity, respectively. In that case, the extracting unit 320 extracts a recommended condition to which all elements of the information indicated by the detection environment match.
  • elements that are not particularly defined in the recommended conditions are regarded as having no restrictions.
  • the selection unit 340 selects recommended combination information indicating a combination having a particularly high usefulness from the recommended combination information associated with the recommended condition extracted by the extraction unit 320.
  • the selection unit 340 can select one or more pieces of recommended combination information from the recommended combination information associated with the extracted recommended conditions based on the prediction accuracy. Specifically, in the recommended information, the prediction accuracy of the odor component when the recommended condition and the recommended combination information use the combination indicated by the recommended combination information may be further associated with the recommended condition and the recommended combination information. Then, the selecting unit 340 can select the recommended combination information having the highest prediction accuracy from the recommended combination information associated with the extracted recommended condition.
  • the selection unit 340 may select one or more pieces of recommended combination information based on the cost when the combination indicated by the recommended combination information is used. Specifically, in the recommended information, the evaluation result based on at least the cost may be further associated with the recommended condition and the recommended combination information. Then, the selecting unit 340 can select the recommended combination information having the best evaluation result among the recommended combination information associated with the extracted recommended conditions.
  • the selection unit 340 may select one or more pieces of recommended combination information based on, for example, the number of sensors 10 included in the combination indicated by the recommended combination information. Specifically, in the recommended information, an evaluation result based on at least the number of the sensors 10 may be further associated with the recommended condition and the recommended combination information. Then, the selecting unit 340 can select the recommended combination information having the best evaluation result among the recommended combination information associated with the extracted recommended conditions.
  • the evaluation result based on at least one of the cost and the number of the sensors 10 may be calculated by the selection unit 340.
  • the selection unit 340 can calculate the evaluation result by the same method as that performed by the evaluation unit 240 of the information processing device 20 in the evaluation step S240. Then, the selection unit 340 selects the recommended combination information based on the calculated evaluation result.
  • the output unit 370 outputs the selected recommended combination information to the selection unit 340.
  • the output unit 370 may store the recommended combination information in a storage device accessible from the output unit 370, may output the recommended combination information to an external device, or may output the recommended combination information to a display device connected to the processing device 30. It may be displayed. The user of the processing device 30 can determine the combination of the sensors 10 used for measurement based on the output recommended combination information.
  • the output unit 370 may further output information indicating a prediction formula in association with the recommended combination information.
  • the information indicating the prediction formula is included in the recommended information, for example, in a state associated with the recommended combination information. The user can make a prediction regarding the odor component using the output prediction formula according to the combination of the sensors 10.
  • the selection unit 340 may select recommended combination information whose prediction accuracy is higher than a predetermined reference instead of selecting one recommended combination information having the highest prediction accuracy. Further, instead of selecting one recommended combination information having the best evaluation result, the selection means 340 may select recommended combination information having the evaluation result better than a predetermined reference. In these cases, the output unit 370 outputs all the selected recommended combination information. At this time, one recommended combination information with the highest prediction accuracy or one recommended combination information with the best evaluation result may be output in a state that can be distinguished from other recommended combination information.
  • the processing device 30 may not include the selection unit 340.
  • the output unit 370 outputs all the recommended combination information associated with the extracted recommended condition to the extraction unit 320.
  • the output unit 370 determines that there is no appropriate combination. Outputs the indicated information.
  • the processing device 30 can be realized by a computer 1000 as shown in FIG.
  • the storage device 1080 stores a program module that implements each functional component of the processing device 30.
  • the processing device 30 may be realized by the same computer as that used to realize the information processing device 20, or may be realized by a different computer.
  • the present embodiment it is possible to know a combination of the sensors 10 suitable for a specific detection environment using the recommended information. Consequently, the detection of the odor component and the prediction based on the detection result can be performed with high accuracy.
  • FIG. 12 is a diagram illustrating a configuration of a processing device 40 according to the fourth embodiment.
  • the processing device 40 according to the present embodiment includes recommended information in which the recommended conditions of the detection environment of the sensor 10 and recommended combination information indicating a combination of one or more sensors 10 are associated with each other, and information indicating the available sensors 10. Output the recommended conditions based on this.
  • the recommendation information is information based on an execution result of machine learning that inputs a plurality of feature amounts based on outputs from a set 100 of a plurality of types of sensors 10 and correct answer data.
  • the processing device 40 includes an extraction unit 420, a selection unit 440, and an output unit 470.
  • the extracting unit 420 extracts, from the recommended combination information included in the recommended information, recommended combination information indicating a feasible combination with the sensor 10 included in the information indicating the available sensor 10.
  • the selection unit 440 selects one or more recommended conditions from the recommended conditions associated with the extracted recommended combination information. Then, the output unit 470 outputs the selected recommended condition.
  • FIG. 13 is a flowchart illustrating a processing method according to the fourth embodiment.
  • the recommended condition is determined based on the recommended information that associates the recommended condition of the sensor detection environment with the recommended combination information indicating a combination of one or more sensors 10 and the information indicating the available sensors 10. Is output.
  • the processing method includes an extraction step S420, a selection step S440, and an output step S470.
  • the extraction step S420 the recommended combination information indicating a feasible combination by the sensor 10 included in the information indicating the usable sensor 10 is extracted from the recommended combination information included in the recommended information.
  • the selection step S440 one or more recommended conditions are selected from the recommended conditions associated with the extracted recommended combination information. Then, in the output step S470, the selected recommended condition is output.
  • the recommended information according to the present embodiment is the same as the recommended information generated by the recommended information generating unit 270 according to at least one of the first and second embodiments, for example. Further, in the following description, the information processing device 20 is the same as the information processing device 20 according to at least one of the first and second embodiments. According to the processing device 40 and the processing method according to the present embodiment, it is possible to use the recommendation information to obtain a preferable detection environment condition when a specific combination of sensors 10 is used. This will be described in detail below.
  • the extraction means 420 acquires the recommended information and the information indicating the usable sensor 10.
  • the extracting unit 420 can obtain the recommended information from a storage device or the information processing device 20 accessible from the extracting unit 420. Further, the user of the processing device 40 can input information indicating the sensors 10 that can be used for the processing device 40. Then, the extracting unit 420 acquires the input information indicating the usable sensor 10.
  • the information indicating the usable sensors 10 may indicate a plurality of sensors 10.
  • the information indicating the usable sensors 10 indicates, for example, a combination of the sensors 10 that can be simultaneously used by a user in a sensor module or the like. When the information indicating the usable sensors 10 includes a plurality of sensors 10, the types of these sensors 10 are different from each other. That is, the functional films and the like of these sensors 10 are different from each other.
  • the recommended information includes one or more recommended combination information and a recommended condition associated therewith.
  • the extracting unit 420 extracts all pieces of recommended combination information indicating combinations that can be realized by the information indicating the available sensors 10 from the recommended information.
  • the combination indicated by the extracted recommended combination information may further include another sensor 10 in addition to the sensor 10 indicated by the information indicating the usable sensor 10.
  • the selection unit 440 selects a recommended condition having a particularly high usefulness from among the recommended conditions associated with the recommended combination information extracted by the extraction unit 420.
  • the selection unit 440 can select one or more recommended conditions based on the prediction accuracy. Specifically, in the recommended information, the recommended condition and the recommended combination information may be further associated with the prediction accuracy of the odor component when the recommended condition and the combination indicated by the recommended combination information are used. Then, the selection unit 440 can select a recommended condition having the highest prediction accuracy from among the recommended conditions associated with the extracted recommended combination information.
  • the selection unit 440 may select one or more recommended conditions based on, for example, at least one of the size of the recommended condition and the proximity to a predetermined condition (standard value). Specifically, in the recommended information, an evaluation result based on at least one of the size of the recommended condition and the proximity to a predetermined condition may be further associated with the recommended condition and the recommended combination information. Then, the selecting unit 440 can select the recommended condition having the best evaluation result among the recommended conditions associated with the extracted recommended combination information.
  • the evaluation result based on at least one of the size of the recommended condition and the proximity to the predetermined condition may be calculated by the selection unit 440.
  • the selection unit 440 can calculate the evaluation result by the same method as that performed by the evaluation unit 240 of the information processing device 20 in the evaluation step S240. Then, the selection unit 440 selects a recommended condition based on the calculated evaluation result.
  • the output unit 470 outputs the selected recommended condition to the selection unit 440.
  • the output unit 470 may, for example, store the recommended conditions in a storage device accessible from the output unit 470, output the recommended conditions to an external device, or display the recommended conditions on a display device connected to the processing device 40. You may let it.
  • the user of the processing device 40 can determine the detection environment at the time of measurement based on the output recommended conditions. Then, for example, the user adjusts temperature, humidity, and the like in the measurement so as to realize the determined detection environment.
  • the output unit 470 may further output information indicating a prediction formula in association with the recommended condition.
  • Information indicating the prediction formula is included in the recommended information, for example, in a state associated with the recommended condition. The user can predict the odor component using the output prediction formula according to the detection environment.
  • the selection unit 440 may select a recommended condition whose prediction accuracy is higher than a predetermined reference. In addition, instead of selecting one recommended condition having the best evaluation result, the selection unit 440 may select a recommended condition whose evaluation result is better than a predetermined reference. In these cases, the output unit 470 outputs all the selected recommended conditions. At this time, one recommended condition with the highest prediction accuracy or one recommended condition with the best evaluation result may be output in a state that can be distinguished from other recommended conditions.
  • the processing device 40 may not include the selection unit 440.
  • the output unit 470 outputs all the recommended conditions associated with the extracted recommended combination information to the extraction unit 420.
  • the output unit 470 determines that there is no appropriate condition. Outputs the indicated information.
  • the processing device 40 can be realized by a computer 1000 as shown in FIG.
  • the storage device 1080 stores a program module that implements each functional component of the processing device 40.
  • the processing device 40 may be realized by the same computer as the computer used to realize the information processing device 20, or may be realized by a different computer.
  • the present embodiment it is possible to know the conditions of the detection environment preferable for the sensor 10 to be used, using the recommended information. As a result, detection of the odor component and prediction based on the detection result can be performed with high accuracy according to the condition.
  • a prediction formula generating means for generating a formula using the plurality of feature amounts as variables, and a prediction formula for performing prediction regarding an odor component
  • Extracting means for extracting one or more of the sensors from the set based on a plurality of weights for the plurality of feature amounts in the prediction formula, and generating recommended combination information indicating the combination of the extracted sensors.
  • the extracting means in the prediction formula, satisfying a predetermined condition among the plurality of weights, or extracting the sensor as an output source of the feature amount weighted by the weight that does not satisfy,
  • the information processing device wherein the recommended information includes one or more pieces of the recommended combination information.
  • the prediction formula generation means includes: Using the model including a branch based on the detection environment to generate the prediction formula, An information processing apparatus for associating the recommended combination information with a condition of the detection environment suitable for the prediction formula, the condition of the detection environment based on the condition of the branch, as the recommended condition. 4. 3.
  • the machine learning is a heterogeneous mixture learning further inputting the detection environment associated with the feature amount, The information processing device according to claim 1, wherein the branch condition is generated by the heterogeneous learning. 5. 2. From 4.
  • the information processing device includes at least one of a temperature, a humidity, an atmospheric pressure, a type of a contaminant gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor. . 6. 2. To 5. In the information processing apparatus according to any one of the above, The apparatus further includes a prediction accuracy calculation unit that calculates prediction accuracy of the prediction expression, The information processing apparatus, wherein the recommendation information further includes the prediction accuracy of the prediction formula associated with the recommendation combination information. 7. 2. From 6.
  • the recommendation information further includes an evaluation result of the evaluation unit, which is associated with the recommendation combination information.
  • the evaluation unit evaluates the combination based on the recommended condition associated with the recommended combination information.
  • the processing device Based on information indicating the detection environment and the recommended information relating the recommended conditions of the sensor detection environment and the recommended combination information indicating the combination of one or more sensors, and outputting the combination,
  • the processing device wherein the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data.
  • 12. 11 In the processing apparatus described in the above, From the recommended conditions included in the recommended information, extraction means for extracting the recommended conditions to which the information indicating the detection environment is suitable, From the recommended combination information associated with the extracted recommended conditions, selecting means for selecting one or more of the recommended combination information, Output means for outputting the combination indicated by the selected recommended combination information. 13.
  • the processing device In the processing apparatus described in the above, In the recommended information, the recommended conditions and the recommended combination information, when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated, The processing device, wherein the selection unit selects one or more pieces of the recommended combination information based on the prediction accuracy from the extracted recommended combination information associated with the extracted recommended condition. 14. 12. Or 13. In the processing apparatus described in the above, The processing device, wherein the selection unit selects one or more pieces of the recommended combination information based on a cost when the combination indicated by the recommended combination information is used. 15. Based on information indicating the recommended conditions of the detection environment of the sensor and recommended combination information indicating a combination of one or more sensors, and information indicating the available sensors, the recommended conditions are output.
  • the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data. 16. 15. In the processing apparatus described in the above, From the recommended combination information included in the recommended information, the sensor included in the information indicating the available sensor, the extraction unit that extracts the recommended combination information indicating the feasible combination, From the recommended conditions associated with the extracted recommended combination information, selecting means for selecting one or more of the recommended conditions, Output means for outputting the selected recommended condition. 17. 16.
  • the processing device In the processing apparatus described in the above, In the recommended information, the recommended conditions and the recommended combination information, when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated, The processing device, wherein the selecting unit selects one or more of the recommended conditions based on the prediction accuracy. 18. 16. Or 17. In the processing apparatus described in the above, The processing device, wherein the selecting unit selects one or more of the recommended conditions based on at least one of a size of the recommended condition and a proximity to a predetermined condition. 19. 11. From 18.
  • the detection environment includes at least one of a temperature, a humidity, an atmospheric pressure, a type of an impurity gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor.
  • the detection environment includes at least one of a temperature, a humidity, an atmospheric pressure, a type of an impurity gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor.
  • a prediction formula generating step of generating a prediction formula for performing a prediction on an odor component wherein the prediction formula generation step is a formula using the plurality of feature amounts as variables, Extracting one or more of the sensors from the set based on a plurality of weights for the plurality of feature amounts in the prediction formula, and generating recommended combination information indicating the combination of the extracted sensors.
  • the recommended information includes one or more pieces of the recommended combination information. 22. 21.
  • the prediction formula generation step Using the model including a branch based on the detection environment to generate the prediction formula, An information processing method for associating the recommended combination information with a condition of the detection environment suitable for the prediction equation, the condition of the detection environment based on the condition of the branch, as the recommended condition. 23. 22.
  • the machine learning is a heterogeneous mixture learning further inputting the detection environment associated with the feature amount, The information processing method according to claim 1, wherein the branch condition is generated by the heterogeneous mixture learning. 24. 21. To 23.
  • the information processing method includes at least one of a temperature, a humidity, an atmospheric pressure, a type of an impurity gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor. . 25. 21. To 24.
  • the method further includes a prediction accuracy calculation step of calculating the prediction accuracy of the prediction expression,
  • the information processing method, wherein the recommendation information further includes the prediction accuracy of the prediction formula associated with the recommendation combination information. 26. 21. To 25.
  • the recommendation information further includes an evaluation result in the evaluation step, which is associated with the recommendation combination information. 27. 26.
  • the evaluating step an information processing method for evaluating the combination based on the recommended condition associated with the recommended combination information. 28. Based on the recommended conditions of the sensor detection environment and the recommended information associated with the recommended combination information indicating a combination of one or more sensors, and the information indicating the detection environment, the combination is output.
  • a processing method wherein the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data. 29. 28.
  • an extraction step of extracting the recommended conditions to which information indicating the detection environment is suitable A selection step of selecting one or more of the recommended combination information from the recommended combination information associated with the extracted recommended condition, Outputting the combination indicated by the selected recommended combination information.
  • the recommended conditions and the recommended combination information when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated,
  • a processing method for selecting one or more pieces of the recommended combination information based on a cost when the combination indicated by the recommended combination information is used. 32. Based on information indicating the recommended conditions of the detection environment of the sensor and recommended combination information indicating a combination of one or more sensors, and information indicating the available sensors, the recommended conditions are output.
  • the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data.
  • the sensor included in the information indicating the available sensor an extraction step of extracting the recommended combination information indicating the feasible combination, A selecting step of selecting one or more of the recommended conditions from the recommended conditions associated with the extracted recommended combination information, Outputting the selected recommended condition. 34. 33.
  • the processing method described in the recommended information when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated,
  • the detection environment includes at least one of a temperature, a humidity, an atmospheric pressure, a type of an impurity gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor.
  • the detection environment includes at least one of a temperature, a humidity, an atmospheric pressure, a type of an impurity gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Investigating Or Analyzing Materials By The Use Of Fluid Adsorption Or Reactions (AREA)
  • Testing Or Calibration Of Command Recording Devices (AREA)

Abstract

An information processing device (20) is provided with a recommendation information generation means (270). The recommendation information generation means (270) generates recommendation information on the basis of the execution result of machine learning which uses, as inputs, correct answer data and multiple feature amounts based on outputs from a collection of sensors of multiple types. The recommendation information is obtained by associating the recommendation condition for a sensor detection environment with a combination of one or more sensors. The information processing device (20) is further provided with, for example, a prediction expression generation means (210) and an extraction means (220). The prediction expression generation means (210) generates, through machine learning, a prediction expression which is for predicting an odor component, and in which multiple feature amounts are used as variables. The extraction means (220) extracts one or more sensors from the collection on the basis of multiple weights with respect to the feature amounts in the prediction expression.

Description

情報処理装置、処理装置、情報処理方法、処理方法、決定方法、およびプログラムInformation processing apparatus, processing apparatus, information processing method, processing method, determination method, and program
 本発明は情報処理装置、処理装置、情報処理方法、処理方法、決定方法、およびプログラムに関する。 The present invention relates to an information processing device, a processing device, an information processing method, a processing method, a determining method, and a program.
 ガスをセンサで測定することにより、ガスに関する情報を得る技術が開発されている。 技術 Technology has been developed to obtain information about gas by measuring gas with a sensor.
 特許文献1は、複数のセンサ素子を設けた匂いセンサを開示している。具体的には、複数のセンサ素子にはそれぞれ異なる特性を有する物質吸着膜が設けられており、各センサ素子は作用させようとする分子に特異的な反応を示す構成をとれることが開示されている。 Patent Document 1 discloses an odor sensor provided with a plurality of sensor elements. Specifically, it is disclosed that a plurality of sensor elements are provided with a substance adsorption film having different characteristics, and each sensor element can take a configuration that exhibits a specific reaction to a molecule to be acted on. I have.
国際公開第2017/085939号WO 2017/085939
 しかし、特許文献1には、検出の目的に応じてセンサ素子の組み合わせをどのように選定すればよいかや、好ましい検出環境について開示されていない。 However, Patent Document 1 does not disclose how to select a combination of sensor elements according to the purpose of detection or a preferable detection environment.
 本発明は、上記の課題に鑑みてなされたものである。本発明の目的は、所望の目的のために、適したセンサの組み合わせ、または好ましい検出環境を導出する技術を提供することにある。 The present invention has been made in view of the above problems. An object of the present invention is to provide a technique for deriving a suitable combination of sensors or a preferable detection environment for a desired purpose.
 本発明の情報処理装置は、
 複数種類のセンサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づき、前記センサの検出環境の推奨条件と、一以上の前記センサからなる組み合わせとを関連づけた推奨情報を生成する推奨情報生成手段を備える。
The information processing device of the present invention includes:
Based on execution results of machine learning using a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data, based on a recommended condition of a detection environment of the sensor and a combination of one or more of the sensors And recommendation information generating means for generating recommendation information in which the information is associated with.
 本発明の第1の決定方法は、
 本発明の情報処理装置で生成された前記推奨情報と、前記検出環境を示す情報とに基づいて、使用する前記センサを決定する方法である。
The first determination method of the present invention is as follows:
A method for determining the sensor to be used based on the recommendation information generated by the information processing apparatus of the present invention and information indicating the detection environment.
 本発明の第2の決定方法は、
 本発明の情報処理装置で生成された前記推奨情報と、使用可能な前記センサを示す情報とに基づいて、前記検出環境を決定する方法である。
The second determination method of the present invention includes:
This is a method for determining the detection environment based on the recommendation information generated by the information processing apparatus of the present invention and information indicating the available sensors.
 本発明の第1の処理装置は、
 センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、前記検出環境を示す情報とに基づいて、前記組み合わせを出力し、
 前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である。
The first processing device of the present invention comprises:
Based on information indicating the detection environment and the recommended information relating the recommended conditions of the sensor detection environment and the recommended combination information indicating the combination of one or more sensors, and outputting the combination,
The recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data.
 本発明の第2の処理装置は、
 センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、使用可能な前記センサを示す情報とに基づいて、前記推奨条件を出力し、
 前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である。
The second processing device of the present invention comprises:
Based on information indicating the recommended conditions of the detection environment of the sensor and recommended combination information indicating a combination of one or more sensors, and information indicating the available sensors, the recommended conditions are output.
The recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data.
 本発明の情報処理方法は、
 複数種類のセンサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づき、前記センサの検出環境の推奨条件と、一以上の前記センサからなる組み合わせとを関連づけた推奨情報を生成する推奨情報生成ステップを含む。
The information processing method of the present invention includes:
Based on execution results of machine learning using a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data, based on a recommended condition of a detection environment of the sensor and a combination of one or more of the sensors And generating recommended information in which recommended information is associated with.
 本発明の第1の処理方法は、
 センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、前記検出環境を示す情報とに基づいて、前記組み合わせを出力し、
 前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である。
The first processing method of the present invention is as follows.
Based on information indicating the detection environment and the recommended information relating the recommended conditions of the sensor detection environment and the recommended combination information indicating the combination of one or more sensors, and outputting the combination,
The recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data.
 本発明の第2の処理方法は、
 センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、使用可能な前記センサを示す情報とに基づいて、前記推奨条件を出力し、
 前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である。
The second processing method of the present invention includes:
Based on information indicating the recommended conditions of the detection environment of the sensor and recommended combination information indicating a combination of one or more sensors, and information indicating the available sensors, the recommended conditions are output.
The recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data.
 本発明の第1のプログラムは、
 本発明の情報処理方法の各ステップをコンピュータに実行させる。
The first program of the present invention is:
A computer is caused to execute each step of the information processing method of the present invention.
 本発明の第2のプログラムは、
 本発明の処理方法の各ステップをコンピュータに実行させる。
A second program according to the present invention includes:
The computer is caused to execute each step of the processing method of the present invention.
 本発明によれば、所望の目的のために、適したセンサの組み合わせ、または好ましい検出環境を導出する技術を提供できる。 According to the present invention, it is possible to provide a technique for deriving a suitable combination of sensors or a preferable detection environment for a desired purpose.
 上述した目的、およびその他の目的、特徴および利点は、以下に述べる好適な実施の形態、およびそれに付随する以下の図面によってさらに明らかになる。 The above and other objects, features and advantages will become more apparent from the preferred embodiments described below and the accompanying drawings.
第1の実施形態に係る情報処理装置の構成を例示する図である。FIG. 2 is a diagram illustrating a configuration of the information processing apparatus according to the first embodiment. センサを例示する図である。It is a figure which illustrates a sensor. 時系列データを例示する図である。It is a figure which illustrates time series data. 複数種類のセンサの集合からのセンサ出力データを例示する図である。It is a figure which illustrates sensor output data from a set of a plurality of types of sensors. 第1の実施形態に係る情報処理方法を例示するフローチャートである。5 is a flowchart illustrating an information processing method according to the first embodiment. 第1の実施形態に係る予測式生成手段で行われる機械学習に用いられる予測モデルを例示する図である。FIG. 3 is a diagram illustrating a prediction model used for machine learning performed by a prediction formula generation unit according to the first embodiment. 情報処理装置を実現するための計算機を例示する図である。FIG. 2 is a diagram illustrating a computer for realizing an information processing device. 第2の実施形態に係る情報処理装置の構成を例示する図である。FIG. 9 is a diagram illustrating a configuration of an information processing apparatus according to a second embodiment. 第2の実施形態に係る情報処理方法を例示するフローチャートである。9 is a flowchart illustrating an information processing method according to a second embodiment. 第3の実施形態に係る処理装置の構成を例示する図である。It is a figure which illustrates the composition of the processing device concerning a 3rd embodiment. 第3の実施形態に係る処理方法を例示するフローチャートである。13 is a flowchart illustrating a processing method according to a third embodiment. 第4の実施形態に係る処理装置の構成を例示する図である。It is a figure which illustrates the composition of the processing device concerning a 4th embodiment. 第4の実施形態に係る処理方法を例示するフローチャートである。13 is a flowchart illustrating a processing method according to a fourth embodiment.
 以下、本発明の実施の形態について、図面を用いて説明する。尚、すべての図面において、同様な構成要素には同様の符号を付し、適宜説明を省略する。 Hereinafter, embodiments of the present invention will be described with reference to the drawings. In all the drawings, the same components are denoted by the same reference numerals, and description thereof will not be repeated.
 なお、以下に示す説明において、特に説明する場合を除き、各装置の各構成要素は、ハードウエア単位の構成ではなく、機能単位のブロックを示している。各装置の各構成要素は、任意のコンピュータのCPU、メモリ、メモリにロードされた本図の構成要素を実現するプログラム、そのプログラムを格納するハードディスクなどの記憶メディア、ネットワーク接続用インタフェースを中心にハードウエアとソフトウエアの任意の組合せによって実現される。そして、その実現方法、装置には様々な変形例がある。 In the following description, each component of each device is not a configuration of a hardware unit but a block of a functional unit, unless otherwise specified. Each component of each device is composed mainly of a CPU of an arbitrary computer, a memory, a program for realizing the components of this drawing loaded in the memory, a storage medium such as a hard disk for storing the program, and a network connection interface. It is realized by any combination of software and software. There are various modifications in the method and apparatus for realizing the method.
(第1の実施形態)
 図1は、第1の実施形態に係る情報処理装置20の構成を例示する図である。本実施形態に係る情報処理装置20は、推奨情報生成手段270を備える。推奨情報生成手段270は、複数種類のセンサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づき、推奨情報を生成する。推奨情報は、センサの検出環境の推奨条件と、一以上のセンサからなる組み合わせとを関連づけた情報である。以下に詳しく説明する。
(First embodiment)
FIG. 1 is a diagram illustrating a configuration of an information processing device 20 according to the first embodiment. The information processing device 20 according to the present embodiment includes the recommended information generation unit 270. The recommendation information generating means 270 generates recommendation information based on a result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data. The recommended information is information that associates a recommended condition of a sensor detection environment with a combination of one or more sensors. This will be described in detail below.
 本図の例において、情報処理装置20は、予測式生成手段210および抽出手段220をさらに備える。予測式生成手段210は、機械学習を行うことにより、複数の特徴量を変数とする式であって、におい成分に関する予測を行うための予測式を生成する。抽出手段220は、予測式における複数の特徴量に対する複数の重みに基づいて、集合から一以上のセンサを抽出する。具体的には抽出手段220は、予測式において、複数の重みのうち予め定められた条件を満たす、または満たさない重みで重みづけられた特徴量の、出力元であるセンサを抽出する。そして、抽出手段220は、抽出されたセンサからなる組み合わせを示す推奨組み合わせ情報を生成する。また、推奨情報は一以上の推奨組み合わせ情報を含む。 に お い て In the example of this figure, the information processing apparatus 20 further includes a prediction formula generation unit 210 and an extraction unit 220. The prediction formula generation means 210 generates a prediction formula for predicting an odor component, which is a formula using a plurality of feature amounts as variables by performing machine learning. The extracting unit 220 extracts one or more sensors from the set based on a plurality of weights for a plurality of feature amounts in the prediction formula. More specifically, the extraction unit 220 extracts a sensor that is an output source of a feature amount that is weighted with a weight that satisfies or does not satisfy a predetermined condition among a plurality of weights in the prediction formula. Then, the extracting unit 220 generates recommended combination information indicating the combination of the extracted sensors. The recommended information includes one or more recommended combination information.
 図2は、センサ10を例示する図である。センサ10は、分子が付着する受容体を有し、その受容体における分子の付着と離脱に応じて検出値が変化するセンサである。なお、センサ10によってセンシングされているガスを、対象ガスと呼ぶ。また、センサ10から出力される検出値の時系列データを、時系列データ14と呼ぶ。ここで、必要に応じ、時系列データ14をYとも表記し、時刻tの検出値をy(t)とも表記する。Yは、y(t)が列挙されたベクトルとなる。 FIG. 2 is a diagram illustrating the sensor 10. The sensor 10 has a receptor to which a molecule is attached, and a detection value changes according to attachment and detachment of the molecule at the receptor. Note that the gas sensed by the sensor 10 is referred to as a target gas. The time-series data of the detection values output from the sensor 10 is referred to as time-series data 14. Here, as necessary, the time-series data 14 is also described as Y, and the detected value at the time t is also described as y (t). Y is a vector in which y (t) is enumerated.
 例えばセンサ10は、膜型表面応力センサ(Membrane-type Surface stress Sensor; MSS)である。MSSは、受容体として、分子が付着する官能膜を有しており、その官能膜に対する分子の付着と離脱によってその官能膜の支持部材に生じる応力が変化する。MSSは、この応力の変化に基づく検出値を出力する。 For example, the sensor 10 is a membrane-type surface stress sensor (MSS). The MSS has, as a receptor, a functional film to which a molecule is attached, and the stress generated in a support member of the functional film is changed by attachment and detachment of the molecule to and from the functional film. The MSS outputs a detection value based on the change in the stress.
 MSSの官能膜には有機系、無機系、およびバイオ系のように様々な材料を用いることができる。センサ10の応答する対象分子および、応答特性は官能膜に依存する。したがって、互いに異なる官能膜を有する複数種類のセンサ10を組み合わせることにより、様々な成分を含む混合ガスからなる複雑なにおいを分析可能となる。 官能 Various materials such as organic, inorganic, and bio-based materials can be used for the MSS functional film. The target molecule to which the sensor 10 responds and the response characteristics depend on the functional film. Therefore, by combining a plurality of types of sensors 10 having different functional films from each other, it becomes possible to analyze a complicated odor composed of a mixed gas containing various components.
 なお、センサ10は、MSSには限定されず、受容体に対する分子の付着と離脱に応じて生じる、センサ10の部材の粘弾性や動力学特性(質量や慣性モーメントなど)に関連する物理量の変化に基づいて検出値を出力するものであればよく、カンチレバー式、膜型、光学式、ピエゾ、振動応答などの様々なタイプのセンサを採用することができる。これらのセンサ10においても、センサ10が応答する対象分子および、応答特性の少なくとも一方が互いに異なる複数種類のセンサ10を組み合わせることができる。なお、複数種類のセンサ10は、同じ属性の情報(付着分子の質量等)を検出してよい。 Note that the sensor 10 is not limited to the MSS, and changes in physical quantities related to the viscoelasticity and dynamic characteristics (mass, moment of inertia, etc.) of the members of the sensor 10 that occur in response to attachment and detachment of molecules to and from the receptor. Any type of sensor may be used as long as it outputs a detection value based on the above, and various types of sensors such as a cantilever type, a film type, an optical type, a piezo, and a vibration response can be adopted. Also in these sensors 10, a plurality of types of sensors 10 having different target molecules to which the sensor 10 responds and at least one of the response characteristics can be combined. Note that the plurality of types of sensors 10 may detect information having the same attribute (such as the mass of an attached molecule).
 ここで、センサ10の種類は多数にのぼる。一方で、実際に検出装置において用いることができるセンサ10の数には限りがある。そこで、目的の検出を行うためにどの種類のセンサ10を組み合わせて用いるのがよいかを選定する必要がある。また、センサ10の出力は、その検出環境にも依存する。したがって、環境に依存して、好ましいセンサ10の組み合わせは異なりうる。同様に、用いるセンサ10の組み合わせに依存して、好ましい検出環境の条件は異なりうる。 Here, the types of the sensors 10 are many. On the other hand, the number of sensors 10 that can be actually used in the detection device is limited. Therefore, it is necessary to select which type of sensor 10 should be used in combination to perform the target detection. The output of the sensor 10 also depends on its detection environment. Thus, depending on the environment, preferred sensor 10 combinations may be different. Similarly, depending on the combination of sensors 10 used, the conditions of the preferred detection environment may vary.
 本実施形態に係る情報処理装置20によれば、推奨される検出環境とセンサ10の組み合わせとを関連づけた推奨情報が得られる。そして、推奨情報に基づいて、好ましいセンサ10の組み合わせおよび検出環境での検出が可能となる。 According to the information processing device 20 according to the present embodiment, the recommended information in which the recommended detection environment and the combination of the sensor 10 are associated with each other can be obtained. Then, based on the recommended information, a preferable combination of the sensors 10 and detection in a detection environment can be performed.
 本実施形態において予測式生成手段210は、複数種類のセンサ10の集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習を行うことにより、におい成分に関する予測を行うための予測式を生成する。予測式は複数の特徴量を変数とする式であり、予測式において各特徴量に対する重みは、その特徴量が予測結果に及ぼす寄与の大きさに対応する。したがって、抽出手段220は予測式を示す情報に基づき、目的に対して寄与が大きなセンサ10と寄与が小さなセンサ10とを判別することができる。 In the present embodiment, the prediction formula generation unit 210 performs prediction regarding an odor component by performing machine learning using a plurality of feature amounts based on outputs from a set of a plurality of types of sensors 10 and correct data as inputs. Generate a prediction equation for. The prediction expression is an expression using a plurality of feature amounts as variables, and the weight for each feature amount in the prediction expression corresponds to the magnitude of the contribution of the feature amount to the prediction result. Therefore, the extracting unit 220 can determine the sensor 10 that has a large contribution to the purpose and the sensor 10 that has a small contribution based on the information indicating the prediction formula.
 ここで、予測式生成手段210は、検出環境に基づいた分岐を含むモデルを用いて予測式を生成する。そうすることで、好ましいセンサ10の組み合わせが検出環境に対応付けて導き出される。また、予測式生成手段210は、推奨組み合わせ情報に、予測式に適した検出環境の条件であって、分岐の条件に基づく検出環境の条件を推奨条件として関連づける。そして、このように互いに関連づけられた推奨組み合わせ情報と推奨条件とを複数組含む推奨情報が生成される。 Here, the prediction formula generation means 210 generates a prediction formula using a model including a branch based on the detection environment. By doing so, a preferable combination of the sensors 10 is derived in association with the detection environment. In addition, the prediction formula generation unit 210 associates the recommended combination information with a detection environment condition suitable for the prediction formula, which is a detection environment condition based on a branch condition, as a recommended condition. Then, recommended information including a plurality of sets of the recommended combination information and the recommended condition associated with each other is generated.
 特徴量および予測式について以下に詳しく説明する。特徴量はセンサ10の出力に基づいて得られる値である。ただし、一つのセンサ10に対しては一つ以上の特徴量が得られ、各特徴量は、一つのセンサ10の出力にのみ依存する。 The feature amount and the prediction formula will be described in detail below. The feature amount is a value obtained based on the output of the sensor 10. However, one or more feature amounts are obtained for one sensor 10, and each feature amount depends only on the output of one sensor 10.
 時系列データ14は、センサ10が出力した検出値を、センサ10から出力された時刻が早い順に並べた時系列のデータである。ただし、時系列データ14は、センサ10から得られた検出値の時系列データに対して、所定の前処理が加えられたものであってもよい。前処理としては、例えば、時系列のデータからノイズ成分を除去するフィルタリングなどを採用することができる。 The time-series data 14 is time-series data in which the detection values output by the sensor 10 are arranged in ascending order of the time output from the sensor 10. However, the time-series data 14 may be obtained by subjecting the time-series data of the detection value obtained from the sensor 10 to predetermined preprocessing. As the pre-processing, for example, filtering for removing a noise component from the time-series data can be employed.
 図3は、時系列データ14を例示する図である。時系列データ14は、センサ10を対象ガスに曝すことで得られる。ただし、時系列データ14は、センサ10を測定対象のガスに曝す操作と、センサ10から測定対象のガスを取り除く操作とで得ても良い。本図の例において、センサ10を対象ガスに曝すことで期間P1のデータが得られ、センサ10から測定対象のガスを取り除く操作により期間P2のデータが得られる。なお、センサ10から測定対象のガスを取り除く操作はたとえばセンサ10をパージガスに曝す操作が挙げられる。また、センサ10による対象ガスの測定においては、センサ10を測定対象のガスに曝す操作と、センサ10から測定対象のガスを取り除く操作を繰り返し行い、複数の時系列データ14を得ても良い。 FIG. 3 is a diagram illustrating the time-series data 14. The time-series data 14 is obtained by exposing the sensor 10 to a target gas. However, the time-series data 14 may be obtained by an operation of exposing the sensor 10 to the gas to be measured and an operation of removing the gas to be measured from the sensor 10. In the example of this figure, data of the period P1 is obtained by exposing the sensor 10 to the target gas, and data of the period P2 is obtained by an operation of removing the gas to be measured from the sensor 10. The operation of removing the gas to be measured from the sensor 10 includes, for example, an operation of exposing the sensor 10 to a purge gas. In the measurement of the target gas by the sensor 10, the operation of exposing the sensor 10 to the gas to be measured and the operation of removing the gas to be measured from the sensor 10 may be repeated to obtain a plurality of time-series data 14.
 図4は、複数種類のセンサ10の集合100からのセンサ出力データ16を例示する図である。本図の例において、センサ10の集合100は、第1センサ10a、第2センサ10b、第3センサ10c、および第4センサ10dからなる。たとえば集合100はモジュール化されており、同じ対象ガスに対して同じ検出環境で測定が行われる。センサ10の集合100は、使用可能な多数のセンサ10から任意に選択された複数のセンサ10からなる。センサ出力データ16は、複数種類のセンサ10のそれぞれから得られた時系列データ14を結合したデータである。本図の例において、センサ出力データ16は、第1センサ10a、第2センサ10b、第3センサ10c、および第4センサ10dの時系列データ14を順に並べたものである。 FIG. 4 is a diagram illustrating sensor output data 16 from a set 100 of a plurality of types of sensors 10. In the example of the figure, a set 100 of the sensors 10 includes a first sensor 10a, a second sensor 10b, a third sensor 10c, and a fourth sensor 10d. For example, the set 100 is modularized, and measurement is performed on the same target gas in the same detection environment. The set 100 of sensors 10 comprises a plurality of sensors 10 arbitrarily selected from a large number of available sensors 10. The sensor output data 16 is data obtained by combining the time-series data 14 obtained from each of the plurality of types of sensors 10. In the example of this figure, the sensor output data 16 is obtained by sequentially arranging the time series data 14 of the first sensor 10a, the second sensor 10b, the third sensor 10c, and the fourth sensor 10d.
 センサ出力データ16からは、複数の特徴量が算出できる。ここで、特徴量ベクトルXを、複数の特徴量を要素とするベクトルであるとする。特徴量ベクトルXには、集合100に含まれる複数種類のセンサ10の出力に基づく複数の特徴量x(j=1,2,...,J)が含まれる。なお、xは数値であっても良いしベクトルであってもよい。xがベクトルである場合、xは同一のセンサ10の出力に基づく複数の特徴量を要素とするベクトルである。特徴量xは、たとえば、センサ10の時系列データ14、時系列データ14を微分したデータ、または、後述する寄与値の集合Ξである。予測式生成手段210は時系列データ14またはセンサ出力データ16を取得し、取得したデータに基づいて特徴量を算出することができる。ただし、予測式生成手段210は時系列データ14またはセンサ出力データ16を取得する代わりに、情報処理装置20の外部で導出された特徴量を取得しても良い。 From the sensor output data 16, a plurality of feature amounts can be calculated. Here, it is assumed that the feature amount vector X is a vector having a plurality of feature amounts as elements. The feature amount vector X includes a plurality of feature amounts x j (j = 1, 2,..., J) based on outputs of a plurality of types of sensors 10 included in the set 100. Note that xj may be a numerical value or a vector. If x j is a vector, x j is a vector whose elements a plurality of feature quantity based on the output of the same sensor 10. The feature amount xj is, for example, the time-series data 14 of the sensor 10, data obtained by differentiating the time-series data 14, or a set 寄 与 of contribution values described later. The prediction formula generation unit 210 can acquire the time-series data 14 or the sensor output data 16 and calculate a feature amount based on the acquired data. However, instead of acquiring the time-series data 14 or the sensor output data 16, the prediction formula generation unit 210 may acquire a feature amount derived outside the information processing device 20.
 予測式は特徴量の線形和であり、z=WX+bで表される。ここで、Wはベクトルであり、bは定数である。そして、重みWの各要素は、特徴量ベクトルXの各要素に対する係数である。そして、得られるzが予測結果を示す。予測式は判別に用いられても良いし、回帰予測に用いられても良い。たとえばあるにおい成分の有無の判別に用いられる予測式では、zが予め定められた基準以上である場合、測定対象のガスに検出対象のにおい成分が含まれていると判断し、基準未満である場合、測定対象のガスに検出対象のにおい成分が含まれていないと判断することができる。回帰予測の例としては、飲料等の製品のにおいに基づく製造品質の予測や呼気の測定による体内状態の予測等が挙げられる。 The prediction formula is a linear sum of the feature amounts, and is represented by z = WX + b. Here, W is a vector and b is a constant. Each element of the weight W is a coefficient for each element of the feature amount vector X. Then, the obtained z indicates the prediction result. The prediction formula may be used for discrimination or may be used for regression prediction. For example, in the prediction formula used to determine the presence or absence of a certain odor component, if z is equal to or greater than a predetermined criterion, it is determined that the gas to be measured contains the odor component to be detected and is smaller than the criterion. In this case, it can be determined that the gas to be measured does not contain the odor component to be detected. Examples of the regression prediction include prediction of manufacturing quality based on the smell of a product such as a beverage, and prediction of a body state by measuring breath.
 なお、上記した時系列データ14、センサ出力データ16、特徴量、および予測式の形態は例であり、本実施形態に係る時系列データ14、センサ出力データ16、特徴量、および予測式の形態は上記に限定されない。 Note that the above-described forms of the time-series data 14, the sensor output data 16, the feature amounts, and the prediction formulas are examples, and the time-series data 14, the sensor output data 16, the feature amounts, and the forms of the prediction formulas according to the present embodiment. Is not limited to the above.
 特徴量の一例である寄与値の集合Ξについて以下に説明する。ここで、説明のため、センサ10によるセンシングを以下のようにモデル化する。
(1)センサ10は、K種類の分子を含む対象ガスに曝されている。
(2)対象ガスにおける各分子kの濃度は一定のρである。
(3)センサ10には、合計N個の分子が吸着可能である。
(4)時刻tにおいてセンサ10に付着している分子kの数はn(t)個である。
A set 寄 与 of contribution values, which is an example of a feature value, will be described below. Here, for explanation, the sensing by the sensor 10 is modeled as follows.
(1) The sensor 10 is exposed to a target gas containing K kinds of molecules.
(2) The concentration of each molecule k in the target gas is constant ρ k .
(3) The sensor 10 can adsorb a total of N molecules.
(4) The number of molecules k attached to the sensor 10 at time t is n k (t).
 センサ10に付着している分子kの数n(t)の時間変化は、以下のように定式化できる。
Figure JPOXMLDOC01-appb-M000001
The change over time of the number n k (t) of molecules k attached to the sensor 10 can be formulated as follows.
Figure JPOXMLDOC01-appb-M000001
 式(1)の右辺の第1項と第2項はそれぞれ、単位時間当たりの分子kの増加量(新たにセンサ10に付着する分子kの数)と減少量(センサ10から離脱する分子kの数)を表している。また、αとβはそれぞれ、分子kがセンサ10に付着する速度を表す速度定数と、分子kがセンサ10から離脱する速度を表す速度定数である。 The first and second terms on the right side of the equation (1) are the increase amount (the number of molecules k newly attached to the sensor 10) and the decrease amount (the molecule k detached from the sensor 10) per unit time. Number). Further, α k and β k are a rate constant representing the rate at which the molecule k adheres to the sensor 10 and a rate constant representing the rate at which the molecule k separates from the sensor 10, respectively.
 ここで、濃度ρが一定であるため、上記式(1)から、時刻tにおける分子kの数n(t)は、以下のように定式化できる。
Figure JPOXMLDOC01-appb-M000002
Here, since the concentration ρ k is constant, the number n k (t) of the numerator k at the time t can be formulated from the above equation (1) as follows.
Figure JPOXMLDOC01-appb-M000002
 また、時刻t(初期状態)でセンサ10に分子が付着していないと仮定すれば、n(t)は以下のように表される。
Figure JPOXMLDOC01-appb-M000003
Assuming that no molecules are attached to the sensor 10 at time t 0 (initial state), n k (t) is expressed as follows.
Figure JPOXMLDOC01-appb-M000003
 センサ10の検出値は、対象ガスに含まれる分子によってセンサ10に働く応力によって定まる。そして、複数の分子によってセンサ10に働く応力は、個々の分子に働く応力の線形和で表すことができると考えられる。ただし、分子によって生じる応力は、分子の種類によって異なると考えられる。すなわち、センサ10の検出値に対する分子の寄与は、その分子の種類によって異なると言える。 検 出 The detection value of the sensor 10 is determined by the stress applied to the sensor 10 by molecules contained in the target gas. Then, it is considered that the stress acting on the sensor 10 by a plurality of molecules can be represented by a linear sum of the stress acting on each molecule. However, it is considered that the stress generated by the molecule differs depending on the type of the molecule. That is, it can be said that the contribution of the molecule to the detection value of the sensor 10 differs depending on the type of the molecule.
 そこで、センサ10の検出値y(t)は、以下のように定式化できる。
Figure JPOXMLDOC01-appb-M000004
 ここで、γとξはいずれも、センサ10の検出値に対する分子kの寄与を表す。なお、「立ち上がり」は上記した期間P1に相当し、「立ち下がり」は上記した期間P2に相当する。
Then, the detection value y (t) of the sensor 10 can be formulated as follows.
Figure JPOXMLDOC01-appb-M000004
Here, both γ k and k k represent the contribution of the numerator k to the detection value of the sensor 10. Note that “rising” corresponds to the above-described period P1, and “falling” corresponds to the above-described period P2.
 ここで、対象ガスをセンシングしたセンサ10から得た時系列データ14を上述の式(4)のように分解できれば、対象ガスに含まれる分子の種類や、各種類の分子が対象ガスに含まれる割合を把握することができる。すなわち、式(4)に示す分解によって、対象ガスの特徴を表すデータ(すなわち、対象ガスの特徴量)が得られる。 Here, if the time-series data 14 obtained from the sensor 10 sensing the target gas can be decomposed as in the above equation (4), the types of molecules contained in the target gas and each type of molecule are contained in the target gas. We can grasp ratio. That is, by the decomposition shown in Expression (4), data representing the characteristics of the target gas (that is, the characteristic amount of the target gas) is obtained.
 そこでセンサ10によって出力された時系列データ14は、特徴定数の集合Θ={θ,θ,...,θ}を用いて、以下の式(5)に示すように分解される。なお、特徴定数の集合Θは、予め定められていてもよいし、情報処理装置20によって生成されてもよい。
Figure JPOXMLDOC01-appb-M000005
 ここで、ξは、センサ10の検出値に対する特徴定数θの寄与を表す寄与値である。
Then, the time-series data 14 output by the sensor 10 includes a set of feature constants Θ = {θ 1 , θ 2 ,. . . , Θ m }, it is decomposed as shown in the following equation (5). The set of feature constants Θ may be determined in advance or may be generated by the information processing device 20.
Figure JPOXMLDOC01-appb-M000005
Here, i i is a contribution value representing the contribution of the characteristic constant θ i to the detection value of the sensor 10.
 このような分解により、時系列データ14に対する各特徴定数θの寄与を表す寄与値ξが算出される。寄与値ξの集合Ξを、対象ガスの特徴を表す特徴量とすることができる。寄与値ξの集合は、例えば、ξを列挙した特徴ベクトルΞ=(ξ,ξ,...,ξ)で表される。ただし、対象ガスの特徴量は、必ずベクトルとして表現しなければならないわけではない。 By such decomposition, a contribution value ξ i representing the contribution of each feature constant θ i to the time series data 14 is calculated. The Ξ set of contribution value xi] i, can be a feature quantity representing the feature of the target gas. A set of contribution values i i is represented, for example, by a feature vector Ξ = (ξ 1 , ξ 2 ,..., M m ) listing ξ i . However, the feature quantity of the target gas does not necessarily have to be represented as a vector.
 ここで、特徴定数θとしては、前述した速度定数βや、速度定数の逆数である時定数τを採用することができる。θとしてβとτを使う場合それぞれについて、式(5)は、以下のように表すことができる。 Here, as the characteristic constant θ, the above-mentioned velocity constant β or a time constant τ which is the reciprocal of the velocity constant can be adopted. For each case where β and τ are used as θ, equation (5) can be expressed as follows.
Figure JPOXMLDOC01-appb-M000006
Figure JPOXMLDOC01-appb-M000006
 前述したように、センサ10の検出値に対する分子の寄与は、その分子の種類によって異なると考えられるため、上述した寄与値の集合Ξは、対象ガスに含まれる分子の種類やその混合比率に応じて異なるものになると考えられる。よって、寄与値の集合Ξは、複数種類の分子が混合されているガスを互いに区別することができる情報、すなわちガスの特徴量として利用することができる。 As described above, the contribution of the molecule to the detection value of the sensor 10 is considered to be different depending on the type of the molecule. Therefore, the set 寄 与 of the contribution values described above depends on the type of the molecule contained in the target gas and the mixing ratio thereof. Are likely to be different. Therefore, the set 寄 与 of contribution values can be used as information that can distinguish a gas in which a plurality of types of molecules are mixed, that is, as a feature amount of the gas.
 寄与値の集合Ξを対象ガスの特徴量として利用することには、複数種類の分子を含むガスを扱えるという利点以外の利点もある。まず、ガス同士の類似度合いを容易に把握することができるという利点がある。例えば、対象ガスの特徴量をベクトルで表現すれば、ガス同士の類似度合いを特徴ベクトル間の距離に基づいて容易に把握することができる。 Using the {set of contribution values} as the feature of the target gas has other advantages besides the advantage of being able to handle gases containing multiple types of molecules. First, there is an advantage that the degree of similarity between gases can be easily grasped. For example, if the feature amount of the target gas is represented by a vector, the degree of similarity between the gases can be easily grasped based on the distance between the feature vectors.
 また、寄与値の集合Ξを特徴量とすることには、混合比変化に対して時定数変化や混合比変化についてロバストにすることができるという利点がある。ここでいう「ロバスト性」とは、「測定環境や測定対象が少しだけ変化したとき、得られる特徴量も少しだけ変化する」という性質である。 Using the {set of contribution values} as the feature quantity has the advantage that it is possible to make the time constant change and the change in the mixture ratio robust against the change in the mixture ratio. The “robustness” here is a property that “when the measurement environment or the measurement target slightly changes, the obtained feature amount also slightly changes”.
 混合比変化についてロバストであれば、例えば、2種類のガスを混合させた混合ガスについて、ガスの混合比を徐々に変化させていくと、特徴量も徐々に変化していくことになる。この性質は、式(4)において、寄与値ξがガスの濃度を表すρに比例しているため、濃度の小さな変化が寄与値の小さな変化として現れるということからわかる。 If the change in the mixture ratio is robust, for example, for a mixed gas in which two types of gases are mixed, if the mixture ratio of the gas is gradually changed, the characteristic amount will also gradually change. This property can be seen from the fact that in equation (4), the contribution value ξ k is proportional to ρ k representing the gas concentration, so that a small change in the concentration appears as a small change in the contribution value.
 図5は、第1の実施形態に係る情報処理方法を例示するフローチャートである。本実施形態に係る情報処理方法は、推奨情報生成ステップS270を含む。推奨情報生成ステップS270では、複数種類のセンサ10の集合100からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づき、推奨情報が生成される。 FIG. 5 is a flowchart illustrating an information processing method according to the first embodiment. The information processing method according to the present embodiment includes a recommended information generating step S270. In the recommended information generation step S270, recommended information is generated based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from the set 100 of a plurality of types of sensors 10 and correct data.
 本図の例において、情報処理方法は、予測式生成ステップS210および抽出ステップS220をさらに含む。予測式生成ステップS210では、機械学習を行うことにより、予測式が生成される。抽出ステップS220では、予測式における複数の特徴量に対する複数の重みに基づいて、集合100から一以上のセンサ10が抽出される。以下に詳しく説明する。 に お い て In the example of this drawing, the information processing method further includes a prediction formula generation step S210 and an extraction step S220. In the prediction formula generation step S210, a prediction formula is generated by performing machine learning. In the extraction step S220, one or more sensors 10 are extracted from the set 100 based on a plurality of weights for a plurality of feature amounts in the prediction formula. This will be described in detail below.
 機械学習の入力とする複数の特徴量は、既知の対象ガスをセンサ10の集合100で測定した結果により得られ、たとえば上記した特徴量ベクトルXである。予測式生成手段210は、時系列データ14、センサ出力データ16、または特徴量ベクトルXを取得する。予測式生成手段210は時系列データ14、センサ出力データ16、または特徴量ベクトルXを、予測式生成手段210からアクセス可能な記憶装置から取得しても良いし、情報処理装置20の外部の装置から取得しても良いし、センサ10から取得しても良い。予測式生成手段210が時系列データ14またはセンサ出力データ16を取得する場合、予測式生成手段210はこれらのデータに基づき特徴量ベクトルXを算出する。特徴量ベクトルXは、その場の測定により得られても良いし、予め準備されて記憶装置に保持されていても良い。 複数 The plurality of feature amounts to be input for machine learning are obtained from the result of measuring a known target gas with the set 100 of the sensors 10 and are, for example, the feature amount vector X described above. The prediction formula generation means 210 acquires the time series data 14, the sensor output data 16, or the feature quantity vector X. The prediction formula generation means 210 may acquire the time-series data 14, the sensor output data 16, or the feature vector X from a storage device accessible from the prediction formula generation means 210, or a device external to the information processing apparatus 20. From the sensor 10 or from the sensor 10. When the prediction formula generation unit 210 acquires the time series data 14 or the sensor output data 16, the prediction formula generation unit 210 calculates the feature quantity vector X based on these data. The feature quantity vector X may be obtained by in-situ measurement, or may be prepared in advance and stored in a storage device.
 また、予測式生成手段210はその特徴量ベクトルXに対する正解データを取得する。正解データは、関連づけられた特徴量ベクトルXに対し予測式で得られるべき予測結果を示す情報である。すなわち、正解データは測定した既知の対象ガスに対応する情報である。正解データはユーザにより情報処理装置20に入力されても良いし、予測式生成手段210からアクセス可能な記憶装置に予め特徴量ベクトルX(すなわち複数の特徴量)と関連づけられて記憶されていても良い。 {Circle around (4)} The prediction formula generation unit 210 acquires correct data for the feature vector X. The correct answer data is information indicating a prediction result to be obtained by a prediction formula for the associated feature amount vector X. That is, the correct answer data is information corresponding to the measured known target gas. The correct answer data may be input to the information processing device 20 by the user, or may be stored in advance in the storage device accessible from the prediction formula generation unit 210 in association with the feature amount vector X (that is, a plurality of feature amounts). good.
 さらに予測式生成手段210は、特徴量に関連づけられた検出環境を取得する。この検出環境は、その特徴量の元となった時系列データ14が得られた際の検出環境である。検出環境は特に限定されないが、たとえば温度、湿度、気圧、夾雑ガスの種類、パージガスの種類、におい成分のサンプリング周期、対象物とセンサ10との距離、センサ10の周囲に存在する物体のうち少なくともいずれかを含む。温度、湿度、および気圧はそれぞれセンサ10の周囲の温度、湿度、および気圧であり、具体的にはセンサ10の官能膜を取り巻く雰囲気の温度、湿度、および気圧である。夾雑ガスの種類は、センサ10を対象ガスに曝す操作において、対象のにおい成分と共にセンサ10に供給されるガスの種類である。具体的には夾雑ガスの種類としては、窒素等の不活性ガス、および空気等が挙げられる。パージガスの種類はセンサ10から測定対象のガスを取り除く操作においてセンサ10に供給されるガスである。具体的にはパージガスとしては、窒素等の不活性ガス、および空気等が挙げられる。におい成分のサンプリング周期は、センサ10を測定対象のガスに曝す操作と、センサ10から測定対象のガスを取り除く操作を繰り返し行う場合の繰り返し周期である。対象物とセンサ10との距離は、特定の対象物の周囲にセンサ10を配置して検出を行う場合の、対象物とセンサ10との距離である。センサ10の周囲に存在する物体は、特定の対象物の周囲にセンサ10を配置して検出を行う場合の、対象物の種類である。 Furthermore, the prediction formula generation means 210 acquires a detection environment associated with the feature amount. This detection environment is a detection environment when the time-series data 14 that is the basis of the feature amount is obtained. Although the detection environment is not particularly limited, for example, temperature, humidity, atmospheric pressure, type of impurity gas, type of purge gas, sampling period of odor component, distance between target and sensor 10, at least one of objects existing around sensor 10 Including any. The temperature, the humidity, and the atmospheric pressure are the temperature, the humidity, and the atmospheric pressure around the sensor 10, and more specifically, the temperature, the humidity, and the atmospheric pressure of the atmosphere surrounding the functional film of the sensor 10. The type of the contaminant gas is the type of gas supplied to the sensor 10 together with the target odor component in the operation of exposing the sensor 10 to the target gas. Specifically, examples of the type of impurity gas include an inert gas such as nitrogen, and air. The type of the purge gas is a gas supplied to the sensor 10 in the operation of removing the gas to be measured from the sensor 10. Specifically, examples of the purge gas include an inert gas such as nitrogen, and air. The sampling cycle of the odor component is a repetition cycle when the operation of exposing the sensor 10 to the gas to be measured and the operation of removing the gas to be measured from the sensor 10 are repeatedly performed. The distance between the target object and the sensor 10 is the distance between the target object and the sensor 10 when the sensor 10 is placed around a specific target object to perform detection. The object existing around the sensor 10 is the type of the target when the sensor 10 is placed around a specific target to perform detection.
 予測式生成手段210は、互いに関連づけられた、複数の特徴量と正解データと検出環境とを含む学習用データセットを入力とする機械学習を行う。予測式生成手段210は、学習用データセットを複数用いて機械学習を行うことで、予測式の精度を高めることができる。このような複数の学習用データセットは、上記した様に、センサ10による対象ガスの測定において、センサ10を測定対象のガスに曝す操作と、センサ10から測定対象のガスを取り除く操作を繰り返し行うことで得られる。予測式生成手段210はたとえば、予め定められた学習の反復回数(学習用データセット数)を満たした場合に学習を終了する。なお、複数の学習用データセットは、検出環境が互いに異なる二以上の学習用データセットを含む。そうすることにより、異種混合学習により分岐の条件が適切に導出され、モデルが生成される。 The prediction formula generation unit 210 performs machine learning using a learning data set including a plurality of feature amounts, correct answer data, and a detection environment, which are associated with each other, as inputs. The prediction formula generation unit 210 can improve the accuracy of the prediction formula by performing machine learning using a plurality of learning data sets. As described above, in the plurality of learning data sets, as described above, in the measurement of the target gas by the sensor 10, the operation of exposing the sensor 10 to the gas to be measured and the operation of removing the gas to be measured from the sensor 10 are repeatedly performed. Obtained by: The prediction formula generation means 210 ends the learning when, for example, a predetermined number of learning iterations (the number of learning data sets) is satisfied. Note that the plurality of learning data sets include two or more learning data sets whose detection environments are different from each other. By doing so, the conditions for branching are appropriately derived by heterogeneous learning, and a model is generated.
 なお、機械学習に用いる特徴量は対象ガスに対するセンサ10の応答をシミュレーションして得られたものであってもよい。なお、複数の学習用データセットは互いに検出環境が異なるシミュレーション条件で得られた結果を用いて生成されうる。ただし、同一の検出環境に対し互いに異なる複数のシミュレーション結果が得られる場合には、複数の学習用データセットは互いに同一のシミュレーション条件で得られた結果を含んで生成されてもよい。 特 徴 Note that the feature amount used for machine learning may be obtained by simulating the response of the sensor 10 to the target gas. Note that a plurality of learning data sets can be generated using results obtained under simulation conditions with different detection environments. However, when a plurality of different simulation results are obtained for the same detection environment, a plurality of learning data sets may be generated including the results obtained under the same simulation conditions.
 図6は、本実施形態に係る予測式生成手段210で行われる機械学習に用いられる予測モデルを例示する図である。本実施形態において機械学習は、複数の特徴量と正解データとに加え、特徴量に関連づけられた検出環境をさらに入力とした異種混合学習である。そして、モデルにおける分岐の条件は、異種混合学習により生成される。 FIG. 6 is a diagram illustrating a prediction model used for machine learning performed by the prediction formula generation unit 210 according to the present embodiment. In the present embodiment, the machine learning is a heterogeneous mixture learning in which, in addition to a plurality of feature amounts and correct answer data, a detection environment associated with the feature amounts is further input. The branch condition in the model is generated by heterogeneous learning.
 機械学習に用いられるモデルは、具体的には複数のノードを含んだ階層構造を有する。そして一以上の中間ノードには分岐の条件として分岐式が位置し、最下層のアノードには予測式が位置する。本図において条件A、条件B1および条件B2は分岐の条件であり、式1から式4はそれぞれ予測式である。なお、中間ノードの数やアノードの数等、モデルの具体的な構成は特に限定されない。 モ デ ル Models used for machine learning have a hierarchical structure that includes a plurality of nodes. A branch formula is located at one or more intermediate nodes as a branch condition, and a prediction formula is located at the lowest anode. In the figure, condition A, condition B1 and condition B2 are branch conditions, and equations 1 to 4 are prediction equations, respectively. The specific configuration of the model such as the number of intermediate nodes and the number of anodes is not particularly limited.
 予測式生成手段210は、予測式生成ステップS210において機械学習を行うことで、一以上の予測式および分岐の条件を含む具体的なモデルを生成する。具体的には、予測式生成手段210は予測式を示す情報として重みWおよび定数bを導出する。また、予測式生成手段210はモデルの構成およびモデルに含まれる各分岐の条件を示す情報を導出する。 The prediction formula generation means 210 generates a specific model including one or more prediction formulas and branch conditions by performing machine learning in the prediction formula generation step S210. Specifically, the prediction formula generation unit 210 derives a weight W and a constant b as information indicating the prediction formula. Further, the prediction formula generation unit 210 derives information indicating the configuration of the model and the conditions of each branch included in the model.
 ここで、各予測式には前提となる検出環境の条件が推奨条件として紐づけられる。各予測式は、その予測式に関連づけられた推奨条件を満たす環境下で特に有効である。推奨条件は予測式と同時に生成されるモデルにおける分岐の条件に基づく。詳しくは、推奨条件は生成されたモデルにおいて、スタートからアノードの予測式に至るまでに通る分岐の条件とその判定結果で定められる。たとえば本図の例において、条件Aが「温度>T」であり、条件B2が「湿度>H」である場合、式3に関連づけられる推奨条件は、「温度がT以下であり、かつ湿度がHより高い」である。 Here, the condition of the detection environment which is a premise is linked to each prediction formula as a recommended condition. Each prediction equation is particularly effective in an environment that satisfies the recommended conditions associated with the prediction equation. The recommended condition is based on a branch condition in a model generated simultaneously with the prediction formula. More specifically, the recommended conditions are determined in the generated model by the conditions of the branches that pass from the start to the prediction formula of the anode and the determination results. For example, in the example of this drawing, when the condition A is “temperature> T 1 ” and the condition B 2 is “humidity> H 1 ”, the recommended condition associated with the equation 3 is “the temperature is T 1 or less, and humidity is higher than H 1 ".
 なお、予測式生成ステップS210において、機械学習で用いる分岐の条件を含む具体的なモデルは、機械学習により生成される代わりに、ユーザにより設定されても良い。この場合、機械学習は異種混合学習でなくても良い。 In addition, in the prediction formula generation step S210, a specific model including a branch condition used in machine learning may be set by a user instead of being generated by machine learning. In this case, the machine learning need not be heterogeneous learning.
 また、異種混合学習では、学習の繰り返しの中で、予測式と共に分岐条件が繰り返し更新されうるが、学習の途中の段階で得られたモデルを、以降の学習で固定して用いても良い。 In addition, in heterogeneous mixture learning, a branching condition can be repeatedly updated together with a prediction formula during repetition of learning, but a model obtained at a stage during learning may be fixed and used in subsequent learning.
 次いで、抽出手段220により抽出ステップS220が行われる。抽出ステップS220において抽出手段220は、各予測式における重みと、重みに関する予め定められた条件とに基づいて、その予測式において予測結果への寄与度が高いセンサ10を抽出する。具体的には、抽出手段220は、予測式生成手段210から予測式を示す情報を取得する。そして、予測式を示す情報に示された各センサ10の特徴量に対する重みの大きさを算出する。 Next, the extraction step 220 is performed by the extraction means 220. In the extraction step S220, the extraction unit 220 extracts the sensor 10 having a high contribution to the prediction result in the prediction formula based on the weight in each prediction formula and a predetermined condition regarding the weight. Specifically, the extraction unit 220 acquires information indicating the prediction formula from the prediction formula generation unit 210. Then, the magnitude of the weight for the characteristic amount of each sensor 10 indicated in the information indicating the prediction formula is calculated.
 ここで、予測式z=WX+bにおけるWXを、集合100に含まれる各センサ10の時系列データ14に基づく特徴量x、および特徴量xに対する重みwを用いて、w+w+・・・wと書き換えることができる。なお、wはそれぞれ数値であっても良いしベクトルであってもよい。wがベクトルである場合、wの各要素は、xの要素である各特徴量に対する重みである。そして、重みの大きさは、たとえばwのノルムである。一方、wが数値である場合、重みの大きさはwの絶対値である。 Here, WX in the prediction equation z = WX + b is calculated as w 1 x 1 + w using a feature amount x j based on the time series data 14 of each sensor 10 included in the set 100 and a weight w j for the feature amount x j . 2 x 2 +... W J x J can be rewritten. Note that w j may be a numerical value or a vector. When w j is a vector, each element of w j is a weight for each feature amount that is an element of x j . The magnitude of the weight is, for example, the norm of w j . On the other hand, when w j is a numerical value, the magnitude of the weight is the absolute value of w j .
 抽出手段220はさらに、算出した重みの大きさが予め定められた条件を満たすか否かを判定する。条件を示す情報は抽出手段220からアクセス可能な記憶装置に予め記憶されている。たとえば、条件が「重みの大きさが基準値以上である」等、予測結果への寄与度が高いセンサ10についての条件を示す場合、抽出手段220はこの条件を満たす重みに対応するセンサ10を抽出する。一方、条件が「重みの大きさが基準値以下である」等、予測結果への寄与が低いセンサ10の条件を示す場合、抽出手段220はこの条件を満たさない重みに対応するセンサ10を抽出する。抽出されるセンサ10の数は特に限定されない。そして抽出手段220は、抽出されたセンサ10からなる組み合わせを示す推奨組み合わせ情報を生成する。生成された推奨組み合わせ情報は、予測式を示す情報に関連づけられる。 The extraction means 220 further determines whether or not the magnitude of the calculated weight satisfies a predetermined condition. The information indicating the condition is stored in a storage device accessible from the extracting unit 220 in advance. For example, when the condition indicates a condition for the sensor 10 having a high degree of contribution to the prediction result, such as “the magnitude of the weight is equal to or greater than the reference value”, the extracting unit 220 determines the sensor 10 corresponding to the weight satisfying this condition. Extract. On the other hand, if the condition indicates a condition of the sensor 10 that makes a small contribution to the prediction result, such as “the magnitude of the weight is equal to or less than the reference value”, the extracting unit 220 extracts the sensor 10 corresponding to the weight that does not satisfy the condition. I do. The number of sensors 10 to be extracted is not particularly limited. Then, the extracting unit 220 generates recommended combination information indicating the combination of the extracted sensors 10. The generated recommended combination information is associated with information indicating a prediction formula.
 抽出手段220は、予測式生成手段210で生成された全ての予測式について、同様にして組み合わせ情報を生成する。 The extraction unit 220 similarly generates combination information for all the prediction expressions generated by the prediction expression generation unit 210.
 なお、図3に示したような時系列データ14において、センサ10に対し吸着および離脱する分子に関する情報は、期間P1および期間P2のそれぞれの冒頭で大きく出力が変動する部分に強く反映されると考えられる。したがって、このような冒頭部分のデータに基づく特徴量の重みが大きくなると予測される。そして仮に、期間P1および期間P2のうち定常部分のデータに基づく特徴量の重みが大きい場合、その結果はノイズ等の影響を受けているとも考えられる。これらのことから、抽出手段220は、期間P1および期間P2の一部のデータのみに基づく特徴量に対する重みに基づいて、センサ10を抽出しても良い。具体的には、期間P1および期間P2のそれぞれにおいて、期間のはじめから予め定められた時間後までの間のデータに基づく特徴量に対する重みに基づいて、センサ10を抽出しても良い。 In the time series data 14 as shown in FIG. 3, information on molecules adsorbed and desorbed from the sensor 10 is strongly reflected in portions where the output fluctuates greatly at the beginning of each of the periods P1 and P2. Conceivable. Therefore, it is predicted that the weight of the feature amount based on the data at the beginning will increase. If the weight of the feature amount based on the data of the stationary part in the period P1 and the period P2 is large, it is considered that the result is affected by noise or the like. From these facts, the extracting means 220 may extract the sensor 10 based on the weight for the feature based on only part of the data in the period P1 and the period P2. Specifically, in each of the period P1 and the period P2, the sensor 10 may be extracted based on the weight for the feature amount based on the data from the beginning of the period to after a predetermined time.
 次いで、推奨情報生成手段270が推奨情報生成ステップS270を行う。推奨情報生成ステップS270において推奨情報生成手段270は抽出手段220から、互いに関連づけられた、予測式を示す情報と、推奨条件と、推奨組み合わせ情報とを取得する。そして推奨情報生成手段270はたとえば、取得したこれらの情報を含む推奨情報を生成する。すなわち推奨情報には、推奨組み合わせ情報と、予測式を示す情報と、推奨条件とが互いに関連づけられた状態で含まれる。なお、推奨情報には、少なくとも推奨組み合わせ情報と、推奨条件とが互いに関連づけられた状態で含まれればよい。 Next, the recommended information generating means 270 performs a recommended information generating step S270. In the recommended information generating step S270, the recommended information generating unit 270 acquires, from the extracting unit 220, information indicating a prediction formula, recommended conditions, and recommended combination information, which are associated with each other. Then, the recommended information generating means 270 generates recommended information including the acquired information, for example. That is, the recommended information includes the recommended combination information, the information indicating the prediction formula, and the recommended condition in a state where they are associated with each other. Note that the recommended information only needs to include at least the recommended combination information and the recommended conditions in a state where they are associated with each other.
 本実施形態では、一つの集合100に対する処理において、推奨情報生成手段270が取得し、推奨情報に含ませる推奨組み合わせ情報の数は、予測式生成手段210で生成される予測式の数に依存する。すなわち予測式生成手段210で生成されるモデルの構成に依存する。推奨情報生成手段270は、抽出手段220で生成された全ての推奨組み合わせ情報を取得し、それら全ての推奨組み合わせ情報にはそれぞれ予測式を示す情報と、推奨条件とが関連づけられている。 In the present embodiment, in the processing for one set 100, the number of pieces of recommended combination information acquired by the recommended information generation unit 270 and included in the recommended information depends on the number of prediction formulas generated by the prediction formula generation unit 210. . That is, it depends on the configuration of the model generated by the prediction expression generation means 210. The recommended information generating unit 270 acquires all the recommended combination information generated by the extracting unit 220, and all the recommended combination information is associated with information indicating a prediction formula and a recommended condition.
 生成された推奨情報は推奨情報生成手段270からアクセス可能な記憶装置に保持されても良いし、外部の装置に対して出力されても良いし、表示装置等でユーザに提示されても良い。 The generated recommended information may be stored in a storage device accessible from the recommended information generating unit 270, may be output to an external device, or may be presented to the user on a display device or the like.
 さらに互いに異なる複数の集合100に対し同様の処理を行うことにより、推奨情報に含まれる推奨組み合わせ情報および推奨条件の数を増やすことができる。この場合、推奨情報生成手段270は、推奨情報生成ステップS270において、推奨情報生成手段270からアクセス可能な記憶装置に保持された既存の推奨情報に、抽出手段220から取得した新たな推奨組み合わせ情報および推奨条件等を追加することで新たな推奨情報を生成する。そして、記憶装置の推奨情報を更新する。 Further, by performing the same processing on a plurality of sets 100 different from each other, the number of pieces of recommended combination information and recommended conditions included in the recommended information can be increased. In this case, the recommendation information generation unit 270 adds the new recommendation combination information acquired from the extraction unit 220 to the existing recommendation information held in the storage device accessible from the recommendation information generation unit 270 in the recommendation information generation step S270. New recommended information is generated by adding recommended conditions and the like. Then, the recommended information of the storage device is updated.
 本実施形態によれば、情報処理装置20で生成された推奨情報と、検出環境を示す情報とに基づいて、使用するセンサ10を決定することができる。たとえば、ユーザが特定の検出環境で使用されるセンサモジュールを作製しようとする場合、多数のセンサ10からセンサモジュールに搭載可能な数の範囲でセンサ10を選択する必要がある。そこで、ユーザは、推奨情報の中から対象の検出環境が該当する推奨条件を見つけ出す。そして、その推奨条件に関連づけられた推奨組み合わせ情報に基づいて、センサモジュールに搭載するセンサ10の組み合わせを決定する。 According to the present embodiment, the sensor 10 to be used can be determined based on the recommended information generated by the information processing device 20 and the information indicating the detection environment. For example, when a user intends to produce a sensor module used in a specific detection environment, it is necessary to select the sensor 10 from a large number of sensors 10 within a range that can be mounted on the sensor module. Therefore, the user finds a recommended condition corresponding to the target detection environment from the recommended information. Then, a combination of the sensors 10 mounted on the sensor module is determined based on the recommended combination information associated with the recommended condition.
 また、本実施形態によれば、情報処理装置20で生成された推奨情報と、使用可能なセンサ10を示す情報とに基づいて、検出環境を決定することができる。たとえば、ユーザが特定の一以上のセンサ10を使用可能であるとき、そのセンサ10を用いて目的の予測を行うためにどのような検出環境下で測定を行えばよいかを推奨情報を用いて知ることができる。具体的には、ユーザは、使用可能なセンサ10の組み合わせで実現可能な推奨組み合わせ情報を見つけ出す。そして、ユーザは、その推奨組み合わせ情報に対応づけられた推奨条件の範囲内で測定を行うのが好ましいと分かる。 According to the present embodiment, the detection environment can be determined based on the recommended information generated by the information processing device 20 and the information indicating the available sensors 10. For example, when the user can use one or more specific sensors 10, the recommended information is used to determine under what detection environment the measurement should be performed in order to perform a target prediction using the sensors 10. You can know. Specifically, the user finds recommended combination information that can be realized by a combination of available sensors 10. Then, the user understands that it is preferable to perform the measurement within the range of the recommended condition associated with the recommended combination information.
 さらに、推奨情報に予測式を示す情報が含まれる場合、ユーザは、最終的に採用した推奨組み合わせ情報と推奨条件に関連づけられた予測式を示す情報に従い、におい成分に関する予測を行うことができる。具体的には、におい成分に関する予測において、複数のセンサ10からの出力に基づき特徴量が算出され、その特徴量が予測式に適用される。そして、予測式による算出値に基づき予測結果が得られる。 Furthermore, when the recommended information includes information indicating a prediction formula, the user can make a prediction regarding the odor component in accordance with the finally adopted recommended combination information and the information indicating the prediction formula associated with the recommended condition. Specifically, in the prediction regarding the odor component, a feature amount is calculated based on the outputs from the plurality of sensors 10, and the feature amount is applied to the prediction formula. Then, a prediction result is obtained based on the value calculated by the prediction formula.
 なお、情報処理装置20は予測の目的ごとに複数の推奨情報を生成しても良い。 Note that the information processing device 20 may generate a plurality of pieces of recommended information for each purpose of prediction.
 情報処理装置20の各機能構成部は、各機能構成部を実現するハードウエア(例:ハードワイヤードされた電子回路など)で実現されてもよいし、ハードウエアとソフトウエアとの組み合わせ(例:電子回路とそれを制御するプログラムの組み合わせなど)で実現されてもよい。以下、情報処理装置20の各機能構成部がハードウエアとソフトウエアとの組み合わせで実現される場合について、さらに説明する。 Each functional component of the information processing device 20 may be implemented by hardware (eg, a hard-wired electronic circuit or the like) that implements each functional component, or a combination of hardware and software (eg: Electronic circuit and a program for controlling the same). Hereinafter, a case where each functional component of the information processing apparatus 20 is realized by a combination of hardware and software will be further described.
 図7は、情報処理装置20を実現するための計算機1000を例示する図である。計算機1000は任意の計算機である。例えば計算機1000は、Personal Computer(PC)やサーバマシンなどの据え置き型の計算機である。その他にも例えば、計算機1000は、スマートフォンやタブレット端末などの可搬型の計算機である。計算機1000は、情報処理装置20を実現するために設計された専用の計算機であってもよいし、汎用の計算機であってもよい。 FIG. 7 is a diagram illustrating a computer 1000 for realizing the information processing device 20. The computer 1000 is an arbitrary computer. For example, the computer 1000 is a stationary computer such as a personal computer (PC) or a server machine. In addition, for example, the computer 1000 is a portable computer such as a smartphone or a tablet terminal. The computer 1000 may be a dedicated computer designed to realize the information processing device 20, or may be a general-purpose computer.
 計算機1000は、バス1020、プロセッサ1040、メモリ1060、ストレージデバイス1080、入出力インタフェース1100、及びネットワークインタフェース1120を有する。バス1020は、プロセッサ1040、メモリ1060、ストレージデバイス1080、入出力インタフェース1100、及びネットワークインタフェース1120が、相互にデータを送受信するためのデータ伝送路である。ただし、プロセッサ1040などを互いに接続する方法は、バス接続に限定されない。 The computer 1000 has a bus 1020, a processor 1040, a memory 1060, a storage device 1080, an input / output interface 1100, and a network interface 1120. The bus 1020 is a data transmission path through which the processor 1040, the memory 1060, the storage device 1080, the input / output interface 1100, and the network interface 1120 mutually transmit and receive data. However, a method for connecting the processors 1040 and the like to each other is not limited to a bus connection.
 プロセッサ1040は、CPU(Central Processing Unit)、GPU(Graphics Processing Unit)、FPGA(Field-Programmable Gate Array)などの種々のプロセッサである。メモリ1060は、RAM(Random Access Memory)などを用いて実現される主記憶装置である。ストレージデバイス1080は、ハードディスク、SSD(Solid State Drive)、メモリカード、又は ROM(Read Only Memory)などを用いて実現される補助記憶装置である。 The processor 1040 is various processors such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), and an FPGA (Field-Programmable Gate Array). The memory 1060 is a main storage device realized using a RAM (Random Access Memory) or the like. The storage device 1080 is an auxiliary storage device realized using a hard disk, an SSD (Solid State Drive), a memory card, or a ROM (Read Only Memory).
 入出力インタフェース1100は、計算機1000と入出力デバイスとを接続するためのインタフェースである。例えば入出力インタフェース1100には、キーボードなどの入力装置や、ディスプレイ装置などの出力装置が接続される。その他にも例えば、入出力インタフェース1100には、センサ10が接続される。ただし、センサ10は必ずしも計算機1000と直接接続されている必要はない。例えばセンサ10は、計算機1000と共有している記憶装置に時系列データ14を記憶させてもよい。 The input / output interface 1100 is an interface for connecting the computer 1000 and an input / output device. For example, an input device such as a keyboard and an output device such as a display device are connected to the input / output interface 1100. In addition, for example, the sensor 10 is connected to the input / output interface 1100. However, the sensor 10 does not necessarily need to be directly connected to the computer 1000. For example, the sensor 10 may store the time-series data 14 in a storage device shared with the computer 1000.
 ネットワークインタフェース1120は、計算機1000を通信網に接続するためのインタフェースである。この通信網は、例えば LAN(Local Area Network)や WAN(Wide Area Network)である。ネットワークインタフェース1120が通信網に接続する方法は、無線接続であってもよいし、有線接続であってもよい。 The network interface 1120 is an interface for connecting the computer 1000 to a communication network. The communication network is, for example, a LAN (Local Area Network) or a WAN (Wide Area Network). The method by which the network interface 1120 connects to the communication network may be a wireless connection or a wired connection.
 ストレージデバイス1080は、情報処理装置20の各機能構成部を実現するプログラムモジュールを記憶している。プロセッサ1040は、これら各プログラムモジュールをメモリ1060に読み出して実行することで、各プログラムモジュールに対応する機能を実現する。 The storage device 1080 stores a program module that implements each functional component of the information processing apparatus 20. The processor 1040 realizes a function corresponding to each program module by reading out each of these program modules into the memory 1060 and executing them.
 次に、本実施形態の作用および効果について説明する。本実施形態に係る情報処理装置20によれば、センサ10の組み合わせと、検出環境の条件との好ましい組み合わせを示す推奨情報を生成することができる。したがって、目的に対して適切なセンサ10の組み合わせおよび環境を知ることができる。 Next, the operation and effect of the present embodiment will be described. According to the information processing apparatus 20 according to the present embodiment, it is possible to generate recommendation information indicating a preferable combination of the combination of the sensors 10 and the conditions of the detection environment. Therefore, it is possible to know the combination of the sensors 10 and the environment appropriate for the purpose.
(第2の実施形態)
 図8は、第2の実施形態に係る情報処理装置20の構成を例示する図である。また、図9は、第2の実施形態に係る情報処理方法を例示するフローチャートである。本実施形態に係る情報処理装置20は、以下に説明する点を除いて第1の実施形態に係る情報処理装置20と同じである。
(Second embodiment)
FIG. 8 is a diagram illustrating a configuration of an information processing device 20 according to the second embodiment. FIG. 9 is a flowchart illustrating an information processing method according to the second embodiment. The information processing device 20 according to the present embodiment is the same as the information processing device 20 according to the first embodiment except for the points described below.
 図8の例において情報処理装置20は、予測式の予測精度を算出する予測精度算出手段230、およびセンサ10の組み合わせを評価する評価手段240をさらに備える。また、図9の例において情報処理方法は、予測精度算出ステップS230および評価ステップS240をさらに含む。ただし、本実施形態に係る情報処理装置20は、予測精度算出手段230および評価手段240の少なくとも一方を備えていなくても良い。また、本実施形態に係る情報処理方法は、予測精度算出ステップS230および評価ステップS240の少なくとも一方を含まなくても良い。 に お い て In the example of FIG. 8, the information processing apparatus 20 further includes a prediction accuracy calculation unit 230 that calculates the prediction accuracy of the prediction formula, and an evaluation unit 240 that evaluates the combination of the sensors 10. In the example of FIG. 9, the information processing method further includes a prediction accuracy calculation step S230 and an evaluation step S240. However, the information processing apparatus 20 according to the present embodiment may not include at least one of the prediction accuracy calculation unit 230 and the evaluation unit 240. Further, the information processing method according to the present embodiment may not include at least one of the prediction accuracy calculation step S230 and the evaluation step S240.
 本実施形態の予測式生成ステップS210および抽出ステップS220では、第1の実施形態に係る予測式生成ステップS210および抽出ステップS220とそれぞれ同様の処理が行われる。 予 測 In the prediction formula generation step S210 and the extraction step S220 of the present embodiment, the same processes as the prediction formula generation step S210 and the extraction step S220 of the first embodiment are performed.
 本実施形態に係る情報処理装置20では、抽出ステップS220に次いで、予測精度算出手段230により予測精度算出ステップS230の処理が行われる。なお、予測精度算出ステップS230の処理が行われるタイミングは予測式生成ステップS210の後かつ、後述する評価ステップS240の前である限り、特に限定されない。なお、情報処理装置20が評価手段240を備えない場合には、予測精度算出ステップS230の処理が行われるタイミングは予測式生成ステップS210の後かつ、推奨情報生成ステップS270の前であればよい。 情報 処理 In the information processing apparatus 20 according to the present embodiment, following the extraction step S220, the processing of the prediction accuracy calculation step S230 is performed by the prediction accuracy calculation means 230. The timing at which the processing of the prediction accuracy calculation step S230 is performed is not particularly limited as long as it is after the prediction expression generation step S210 and before the evaluation step S240 described later. When the information processing apparatus 20 does not include the evaluation unit 240, the timing at which the processing of the prediction accuracy calculation step S230 is performed may be after the prediction expression generation step S210 and before the recommendation information generation step S270.
 予測精度算出ステップS230では、予測精度算出手段230が予測式生成手段210で生成された各予測式の予測精度を算出する。予測精度の算出には、学習用データセットと同様のデータセットが評価用データセットとして用いられる。すなわち、評価用データセットは複数の特徴量と正解データと検出環境とを含む。 In the prediction accuracy calculation step S230, the prediction accuracy calculation means 230 calculates the prediction accuracy of each prediction formula generated by the prediction formula generation means 210. To calculate the prediction accuracy, a data set similar to the learning data set is used as the evaluation data set. That is, the evaluation data set includes a plurality of feature amounts, correct answer data, and a detection environment.
 ただし、複数の学習用データセットと複数の評価用データセットには、互いに全く同じデータセットは含まれない。たとえば情報処理装置20の外部または内部で生成された、互いに異なる複数のデータセットのうちの一部を複数の学習用データセットとして用い、残りを複数の評価用データセットとして用いることができる。 However, the plurality of learning data sets and the plurality of evaluation data sets do not include exactly the same data sets. For example, a part of a plurality of different data sets generated outside or inside the information processing device 20 may be used as a plurality of learning data sets, and the rest may be used as a plurality of evaluation data sets.
 予測精度は回帰に基づく予測については回帰精度であり、たとえば最小二乗誤差または平均平方二乗誤差(RMSE)である。また、予測精度は判別に基づく予測については判別精度であり、たとえば適合率、再現率、F値、正答率、またはROC_AUCである。 Prediction accuracy is the regression accuracy for prediction based on regression, for example, least squares error or mean squared error (RMSE). The prediction accuracy is the determination accuracy for prediction based on the determination, and is, for example, a precision, a recall, an F value, a correct answer rate, or ROC_AUC.
 予測精度算出手段230が予測精度を算出する方法の一例について詳しく説明する。予測式生成手段210が学習用データセットを取得または生成するのと同様の方法で、予測精度算出手段230は複数の評価用データセットを取得または生成することができる。予測精度算出手段230は評価用データセットに含まれる特徴量を、精度を評価しようとする予測式に入力することで、予測結果を得る。そして、得られた予測結果と、評価用データセットに含まれる正解データとが一致するか否かを判定する。そして、予測精度算出手段230は複数の評価用データセットについて同様の処理を行い、予測結果と正解データとが一致する確率を、その予測式の予測精度として算出する。算出された予測精度は、その予測式に関連づけられる。 An example of how the prediction accuracy calculation means 230 calculates the prediction accuracy will be described in detail. The prediction accuracy calculation unit 230 can obtain or generate a plurality of evaluation data sets in the same manner as the prediction expression generation unit 210 obtains or generates a learning data set. The prediction accuracy calculation means 230 obtains a prediction result by inputting the feature amount included in the evaluation data set into a prediction expression whose accuracy is to be evaluated. Then, it is determined whether or not the obtained prediction result matches the correct answer data included in the evaluation data set. Then, the prediction accuracy calculation means 230 performs the same processing for a plurality of evaluation data sets, and calculates the probability that the prediction result matches the correct answer data as the prediction accuracy of the prediction formula. The calculated prediction accuracy is associated with the prediction formula.
 複数の評価用データセットは、互いに異なる検出環境での測定結果に基づくものであっても良い。ただし、各予測式について、その予測式に関連づけられた検出環境の条件を満たす環境で得られた評価用データセットのみが予測精度の算出に用いられる。 (4) The plurality of evaluation data sets may be based on measurement results in mutually different detection environments. However, for each prediction formula, only the evaluation data set obtained in an environment that satisfies the condition of the detection environment associated with the prediction formula is used for calculating the prediction accuracy.
 次いで、評価ステップS240の処理が評価手段240により行われる。評価手段240は、センサ10の組み合わせを、たとえばその組み合わせおよび検出環境を採用する場合に用いる予測式の予測精度と、その組み合わせを採用する場合のコストとの少なくとも一方に基づいて評価する。なかでも評価手段240は、推奨組み合わせ情報に示されたセンサ10の組み合わせを採用する場合のコストに少なくとも基づいて、センサ10の組み合わせを評価することが好ましい。 Next, the processing of the evaluation step S240 is performed by the evaluation means 240. The evaluation means 240 evaluates the combination of the sensors 10 based on, for example, at least one of the prediction accuracy of a prediction formula used when adopting the combination and the detection environment and the cost when adopting the combination. In particular, it is preferable that the evaluation unit 240 evaluates the combination of the sensors 10 based at least on the cost when the combination of the sensors 10 indicated in the recommended combination information is adopted.
 コストにはたとえば初期コストおよびランニングコストが含まれる。初期コストとしては、センサ10の製造コストや調達コスト等が挙げられる。また、ランニングコストとしては、管理コスト、センサ10の劣化等に起因する交換コスト、扱いにおける人的手間等が挙げられる。 Cost includes initial cost and running cost, for example. Examples of the initial cost include a manufacturing cost and a procurement cost of the sensor 10. The running costs include management costs, replacement costs caused by deterioration of the sensor 10, and human labor in handling.
 評価手段240によりアクセス可能な記憶装置には、予め各センサ10のコストを示すパラメータが保持されており、評価手段240は、組み合わせに含まれるセンサ10のコストを示すパラメータを記憶装置から取得する。そして、組み合わせに含まれる全てのセンサ10についてのコストを示すパラメータを合算し、合算値を得る。 記憶 A parameter indicating the cost of each sensor 10 is stored in advance in a storage device accessible by the evaluation unit 240, and the evaluation unit 240 acquires a parameter indicating the cost of the sensor 10 included in the combination from the storage device. Then, the parameters indicating the costs for all the sensors 10 included in the combination are added up to obtain a total value.
 また、評価手段240は予測精度算出手段230から、その推奨組み合わせ情報に関連づけられた予測式の予測精度を取得する。 {Circle around (4)} The evaluation unit 240 acquires the prediction accuracy of the prediction formula associated with the recommended combination information from the prediction accuracy calculation unit 230.
 評価手段240はさらに評価関数を用いて組み合わせを評価する。評価関数は一以上の要因に基づき評価値を算出する関数である。具体的には評価関数は、各要因における評価結果を示す評価パラメータの線形和で表される。たとえば要因をコストとした評価パラメータは、上記の様に算出された合算値であり、要因を精度とした評価パラメータは予測精度算出手段230から取得した予測精度である。また、評価関数では、各評価パラメータに対して係数が掛けられ、評価結果に対する要因ごとの重みのバランスがとられたり、評価の方向性が定められたりしている。係数は、評価パラメータの種類毎に定められている。 Evaluation means 240 further evaluates the combination using an evaluation function. The evaluation function is a function that calculates an evaluation value based on one or more factors. Specifically, the evaluation function is represented by a linear sum of evaluation parameters indicating the evaluation result of each factor. For example, the evaluation parameter with the cost as the factor is the total value calculated as described above, and the evaluation parameter with the accuracy as the factor is the prediction accuracy acquired from the prediction accuracy calculation unit 230. In the evaluation function, each evaluation parameter is multiplied by a coefficient to balance the weight for each factor with respect to the evaluation result or to determine the directionality of the evaluation. The coefficient is determined for each type of evaluation parameter.
 評価手段240はたとえば評価関数にコストを示すパラメータの合算値および予測精度を適用することにより、評価結果として評価値を算出する。なお、評価手段240により得られる評価結果は、コストに関する合算値は小さいほど高くなり、予測精度が良いほど高くなる。評価関数を示す情報は評価手段240によりアクセス可能な記憶装置に予め保持されている。算出された評価値は、推奨組み合わせ情報に関連づけられる。 Evaluation means 240 calculates an evaluation value as an evaluation result, for example, by applying the sum of parameters indicating cost and prediction accuracy to the evaluation function. The evaluation result obtained by the evaluation means 240 increases as the sum of the costs decreases, and increases as the prediction accuracy improves. Information indicating the evaluation function is stored in a storage device accessible by the evaluation unit 240 in advance. The calculated evaluation value is associated with the recommended combination information.
 評価手段240は、さらに組み合わせに含まれるセンサ10の数に基づき、センサ10の組み合わせを評価しても良い。たとえば、組み合わせに含まれるセンサ10の数を要因とする場合、たとえば、センサ10の数が、評価関数における評価パラメータとなり得る。なお、評価手段240により得られる評価結果は、組み合わせに含まれるセンサ10の数が少ないほど高くなる。 The evaluation means 240 may further evaluate the combination of the sensors 10 based on the number of the sensors 10 included in the combination. For example, when the number of sensors 10 included in the combination is a factor, for example, the number of sensors 10 can be an evaluation parameter in an evaluation function. Note that the evaluation result obtained by the evaluation unit 240 increases as the number of sensors 10 included in the combination decreases.
 また、評価手段240は、推奨組み合わせ情報に関連づけられた推奨条件にさらに基づいて、組み合わせを評価してもよい。たとえば、推奨条件の広さを要因とする場合、たとえば、推奨条件として示された温度、湿度、気圧、周期、距離等の範囲の幅や、ガスや物体の選択肢の数が、評価関数における評価パラメータとなり得る。また、推奨条件の実用性を要因とする場合、推奨条件として示された温度、湿度、気圧、周期、距離等の範囲の中心値と、予め定められた標準値との距離が評価関数における評価パラメータとなり得る。すなわち、この距離が小さいほど実用性が高いといえる。なお、評価手段240により得られる評価結果は、推奨条件が広いほど高くなり、推奨条件の実用性が高いほど高くなる。 {Evaluation means 240 may evaluate a combination further based on recommendation conditions linked with recommended combination information. For example, when the size of the recommended condition is a factor, for example, the width of the range of the temperature, humidity, atmospheric pressure, cycle, distance, etc. indicated as the recommended condition, or the number of options of gas or object is evaluated by the evaluation function. Can be a parameter. When the practicality of the recommended condition is a factor, the distance between the center value of the range of temperature, humidity, atmospheric pressure, cycle, distance, etc. indicated as the recommended condition and a predetermined standard value is evaluated by the evaluation function. Can be a parameter. That is, it can be said that the shorter the distance, the higher the practicality. Note that the evaluation result obtained by the evaluation means 240 increases as the recommended conditions are wider, and increases as the practicality of the recommended conditions increases.
 次いで、推奨情報生成手段270が推奨情報生成ステップS270を行う。推奨情報生成ステップS270において推奨情報生成手段270は、互いに関連づけられた、予測式を示す情報と、推奨条件と、推奨組み合わせ情報と、評価結果とを取得する。そして推奨情報生成手段270は、取得したこれらの情報を含む推奨情報を生成する。すなわち推奨情報には、推奨組み合わせ情報と、予測式を示す情報と、推奨条件と、評価結果とが互いに関連づけられた状態で含まれる。なお、本例では推奨情報が、推奨組み合わせ情報に関連づけられた、評価手段240の評価結果をさらに含む例について説明しているが、推奨情報は、評価結果に代えて、または評価結果に加えて、推奨組み合わせ情報に関連づけられた予測式の予測精度をさらに含んでもよい。 Next, the recommended information generating means 270 performs a recommended information generating step S270. In the recommended information generating step S270, the recommended information generating means 270 acquires information indicating a prediction formula, recommended conditions, recommended combination information, and an evaluation result, which are associated with each other. Then, the recommended information generating unit 270 generates recommended information including the acquired information. That is, the recommended information includes the recommended combination information, the information indicating the prediction formula, the recommended condition, and the evaluation result in a state where they are associated with each other. Note that, in this example, an example is described in which the recommended information further includes the evaluation result of the evaluation unit 240 associated with the recommended combination information. However, the recommended information is replaced with the evaluation result or in addition to the evaluation result. And the prediction accuracy of the prediction formula associated with the recommended combination information.
 また、推奨情報生成手段270は、予測精度および評価結果の少なくとも一方に基づき、予測式の選択を行い、選択された予測式に関連づけられた情報のみが推奨情報に含まれてもよい。具体的には、推奨情報生成手段270は、予測式生成手段210で生成された予測式のうち、予め定められた基準よりも優れた予測精度の予測式を選択する。または、推奨情報生成手段270は、予測式生成手段210で生成された予測式のうち、予め定められた基準よりも優れた評価結果の予測式を選択する。そして、推奨情報生成手段270は、選択された予測式に関連づけられた情報(推奨組み合わせ情報および推奨条件等)を含む推奨情報を生成する。 (4) The recommended information generating means 270 may select a prediction formula based on at least one of the prediction accuracy and the evaluation result, and only the information associated with the selected prediction formula may be included in the recommended information. Specifically, the recommendation information generation unit 270 selects a prediction expression having a higher prediction accuracy than a predetermined reference from the prediction expressions generated by the prediction expression generation unit 210. Alternatively, the recommendation information generation unit 270 selects a prediction expression of an evaluation result superior to a predetermined criterion from among the prediction expressions generated by the prediction expression generation unit 210. Then, the recommended information generating unit 270 generates recommended information including information (recommended combination information, recommended conditions, and the like) associated with the selected prediction formula.
 本実施形態に係る情報処理装置20も、図7に示したような計算機1000により実現可能である。本実施形態において、ストレージデバイス1080は、情報処理装置20の予測精度算出手段230および評価手段240をそれぞれ実現するプログラムモジュールをさらに記憶している。 情報 処理 The information processing apparatus 20 according to the present embodiment can also be realized by the computer 1000 as shown in FIG. In the present embodiment, the storage device 1080 further stores program modules for realizing the prediction accuracy calculation unit 230 and the evaluation unit 240 of the information processing device 20.
 次に、本実施形態の作用および効果について説明する。本実施形態においては第1の実施形態と同様の作用および効果が得られる。くわえて、予測精度算出手段230で予測式の予測精度が算出されたり、評価手段240による評価が行われたりすることで、複数の推奨組み合わせ情報等の有用性を互いに比較することができる。 Next, the operation and effect of the present embodiment will be described. In the present embodiment, the same operation and effect as those of the first embodiment can be obtained. In addition, the prediction accuracy of the prediction formula is calculated by the prediction accuracy calculation unit 230 or the evaluation by the evaluation unit 240 is performed, so that the usefulness of a plurality of pieces of recommended combination information can be compared with each other.
 (第3の実施形態)
 図10は、第3の実施形態に係る処理装置30の構成を例示する図である。本実施形態に係る処理装置30は、センサ10の検出環境の推奨条件と一以上のセンサ10からなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、検出環境を示す情報とに基づいて、組み合わせを出力する。ここで、推奨情報は、複数種類のセンサ10の集合100からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である。
(Third embodiment)
FIG. 10 is a diagram illustrating a configuration of a processing device 30 according to the third embodiment. The processing device 30 according to the present embodiment, based on recommended information in which the recommended conditions of the detection environment of the sensor 10 and the recommended combination information indicating a combination of one or more sensors 10 are associated, and information indicating the detection environment, Output the combination. Here, the recommendation information is information based on an execution result of machine learning that inputs a plurality of feature amounts based on outputs from a set 100 of a plurality of types of sensors 10 and correct answer data.
 本図の例において、処理装置30は、抽出手段320、選択手段340および出力手段370を備える。抽出手段320は、推奨情報に含まれる推奨条件から、検出環境を示す情報が適合する推奨条件を抽出する。選択手段340は、抽出された推奨条件に関連づけられた推奨組み合わせ情報から、一以上の推奨組み合わせ情報を選択する。そして、出力手段370は、選択された推奨組み合わせ情報が示す組み合わせを出力する。 In the example of the figure, the processing device 30 includes an extraction unit 320, a selection unit 340, and an output unit 370. The extracting unit 320 extracts, from the recommended conditions included in the recommended information, a recommended condition to which the information indicating the detection environment matches. The selection unit 340 selects one or more pieces of recommended combination information from the recommended combination information associated with the extracted recommended conditions. Then, the output unit 370 outputs the combination indicated by the selected recommended combination information.
 図11は、第3の実施形態に係る処理方法を例示するフローチャートである。本処理方法では、センサの検出環境の推奨条件と一以上のセンサ10からなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、検出環境を示す情報とに基づいて、組み合わせが出力される。 FIG. 11 is a flowchart illustrating a processing method according to the third embodiment. In this processing method, the combination is output based on the recommended information in which the recommended conditions of the sensor detection environment and the recommended combination information indicating the combination of one or more sensors 10 are associated with each other and the information indicating the detection environment.
 本図の例において、本実施形態に係る処理方法は、抽出ステップS320、選択ステップS340および出力ステップS370を含む。抽出ステップS320では、推奨情報に含まれる推奨条件から、検出環境を示す情報が適合する推奨条件が抽出される。選択ステップS340では、抽出された推奨条件に関連づけられた推奨組み合わせ情報から、一以上の推奨組み合わせ情報が選択される。そして、出力ステップS370では、選択された推奨組み合わせ情報が示す組み合わせが出力される。本実施形態に係る処理方法は処理装置30により実現される。 In the example of the drawing, the processing method according to the present embodiment includes an extraction step S320, a selection step S340, and an output step S370. In the extraction step S320, a recommended condition to which the information indicating the detection environment matches is extracted from the recommended conditions included in the recommended information. In the selection step S340, one or more recommended combination information is selected from the recommended combination information associated with the extracted recommended condition. Then, in the output step S370, the combination indicated by the selected recommended combination information is output. The processing method according to the present embodiment is realized by the processing device 30.
 本実施形態に係る推奨情報は、たとえば第1および第2の実施形態の少なくともいずれかに係る推奨情報生成手段270で生成される推奨情報と同じである。また、以下の説明において情報処理装置20は、第1および第2の実施形態の少なくともいずれかに係る情報処理装置20と同じである。本実施形態に係る処理装置30および処理方法によれば、推奨情報を用いて、センサ10を使用しようとする検出環境から、好ましいセンサ10の組み合わせを求めることができる。以下に詳しく説明する。 The recommended information according to the present embodiment is the same as the recommended information generated by the recommended information generating unit 270 according to at least one of the first and second embodiments, for example. Further, in the following description, the information processing device 20 is the same as the information processing device 20 according to at least one of the first and second embodiments. According to the processing device 30 and the processing method according to the present embodiment, a preferable combination of the sensors 10 can be obtained from the detection environment in which the sensor 10 is to be used, using the recommended information. This will be described in detail below.
 抽出ステップS320において抽出手段320は、推奨情報および検出環境を示す情報を取得する。抽出手段320は推奨情報を、抽出手段320からアクセス可能な記憶装置または情報処理装置20から取得できる。また、処理装置30のユーザは処理装置30に対して検出環境を示す情報を入力可能である。そして、抽出手段320は入力された検出環境を示す情報を取得する。検出環境を示す情報は、たとえば、ユーザがいずれかのセンサ10を用いて測定を行おうとする際の、検出環境を示す。 In the extraction step S320, the extraction means 320 acquires recommended information and information indicating the detection environment. The extracting unit 320 can obtain the recommended information from a storage device or the information processing device 20 accessible from the extracting unit 320. Further, the user of the processing device 30 can input information indicating the detection environment to the processing device 30. Then, the extracting unit 320 acquires information indicating the input detection environment. The information indicating the detection environment indicates, for example, the detection environment when the user intends to perform measurement using any of the sensors 10.
 第1および第2の実施形態において説明した通り、推奨情報には一以上の推奨組み合わせ情報と、それに関連づけられた推奨条件が含まれる。抽出手段320は、推奨情報の中から、検出環境を示す情報が適合する推奨条件を全て抽出する。なお、推奨条件および検出環境を示す情報はそれぞれ温度および湿度等、複数の要素を含んでも良い。その場合、抽出手段320は、検出環境が示す情報の全ての要素が適合する推奨条件を抽出する。ここで、推奨条件において特に定めがない要素については、制限が無いものとみなされる。 As described in the first and second embodiments, the recommended information includes one or more recommended combination information and a recommended condition associated therewith. The extracting unit 320 extracts, from the recommended information, all the recommended conditions to which the information indicating the detection environment matches. The information indicating the recommended condition and the detection environment may include a plurality of elements such as temperature and humidity, respectively. In that case, the extracting unit 320 extracts a recommended condition to which all elements of the information indicated by the detection environment match. Here, elements that are not particularly defined in the recommended conditions are regarded as having no restrictions.
 次いで選択手段340は選択ステップS340において、抽出手段320で抽出された推奨条件に関連づけられた推奨組み合わせ情報のうち、特に有用性が高い組み合わせを示す推奨組み合わせ情報を選択する。 Next, in the selection step S340, the selection unit 340 selects recommended combination information indicating a combination having a particularly high usefulness from the recommended combination information associated with the recommended condition extracted by the extraction unit 320.
 たとえば、選択手段340は、抽出された推奨条件に関連づけられた推奨組み合わせ情報から、予測精度に基づいて、一以上の推奨組み合わせ情報を選択することができる。具体的には、推奨情報において、推奨条件および推奨組み合わせ情報に、その推奨条件とその推奨組み合わせ情報が示す組み合わせとを用いた場合の、におい成分に関する予測精度がさらに関連づけられていてもよい。すると、選択手段340は、抽出された推奨条件に関連づけられた推奨組み合わせ情報のうち、予測精度が最も高い推奨組み合わせ情報を選択することができる。 {For example, the selection unit 340 can select one or more pieces of recommended combination information from the recommended combination information associated with the extracted recommended conditions based on the prediction accuracy. Specifically, in the recommended information, the prediction accuracy of the odor component when the recommended condition and the recommended combination information use the combination indicated by the recommended combination information may be further associated with the recommended condition and the recommended combination information. Then, the selecting unit 340 can select the recommended combination information having the highest prediction accuracy from the recommended combination information associated with the extracted recommended condition.
 また、選択手段340はたとえば、推奨組み合わせ情報が示す組み合わせを用いた場合の、コストに基づいて、一以上の推奨組み合わせ情報を選択してもよい。具体的には、推奨情報において、推奨条件および推奨組み合わせ情報に、少なくともコストに基づく評価結果がさらに関連づけられていてもよい。すると、選択手段340は、抽出された推奨条件に関連づけられた推奨組み合わせ情報のうち、評価結果が最も優れる推奨組み合わせ情報を選択することができる。 (4) For example, the selection unit 340 may select one or more pieces of recommended combination information based on the cost when the combination indicated by the recommended combination information is used. Specifically, in the recommended information, the evaluation result based on at least the cost may be further associated with the recommended condition and the recommended combination information. Then, the selecting unit 340 can select the recommended combination information having the best evaluation result among the recommended combination information associated with the extracted recommended conditions.
 また、選択手段340はたとえば、推奨組み合わせ情報が示す組み合わせに含まれるセンサ10の数に基づいて、一以上の推奨組み合わせ情報を選択してもよい。具体的には、推奨情報において、推奨条件および推奨組み合わせ情報に、少なくともセンサ10の数に基づく評価結果がさらに関連づけられていてもよい。すると、選択手段340は、抽出された推奨条件に関連づけられた推奨組み合わせ情報のうち、評価結果が最も優れる推奨組み合わせ情報を選択することができる。 The selection unit 340 may select one or more pieces of recommended combination information based on, for example, the number of sensors 10 included in the combination indicated by the recommended combination information. Specifically, in the recommended information, an evaluation result based on at least the number of the sensors 10 may be further associated with the recommended condition and the recommended combination information. Then, the selecting unit 340 can select the recommended combination information having the best evaluation result among the recommended combination information associated with the extracted recommended conditions.
 なお、コストおよびセンサ10の数の少なくともいずれかに基づく評価結果は選択手段340で算出されても良い。この場合、選択手段340は、情報処理装置20の評価手段240が評価ステップS240において行ったのと同様の方法で評価結果を算出することができる。そして、選択手段340は算出した評価結果に基づいて推奨組み合わせ情報を選択する。 The evaluation result based on at least one of the cost and the number of the sensors 10 may be calculated by the selection unit 340. In this case, the selection unit 340 can calculate the evaluation result by the same method as that performed by the evaluation unit 240 of the information processing device 20 in the evaluation step S240. Then, the selection unit 340 selects the recommended combination information based on the calculated evaluation result.
 次いで、出力ステップS370において出力手段370は、選択手段340に選択された推奨組み合わせ情報を出力する。出力手段370はたとえば、推奨組み合わせ情報を出力手段370からアクセス可能な記憶装置に記憶させても良いし、外部の装置に対して出力しても良いし、処理装置30に接続された表示装置に表示させても良い。処理装置30のユーザは、出力された推奨組み合わせ情報に基づき、測定に用いるセンサ10の組み合わせを決定することができる。 Next, in the output step S370, the output unit 370 outputs the selected recommended combination information to the selection unit 340. The output unit 370 may store the recommended combination information in a storage device accessible from the output unit 370, may output the recommended combination information to an external device, or may output the recommended combination information to a display device connected to the processing device 30. It may be displayed. The user of the processing device 30 can determine the combination of the sensors 10 used for measurement based on the output recommended combination information.
 また、出力手段370は、推奨組み合わせ情報に関連づけて、予測式を示す情報をさらに出力しても良い。予測式を示す情報はたとえば推奨組み合わせ情報に関連づけられた状態で推奨情報に含まれる。ユーザは、センサ10の組み合わせに応じて、出力された予測式を用い、におい成分に関する予測を行うことができる。 (4) The output unit 370 may further output information indicating a prediction formula in association with the recommended combination information. The information indicating the prediction formula is included in the recommended information, for example, in a state associated with the recommended combination information. The user can make a prediction regarding the odor component using the output prediction formula according to the combination of the sensors 10.
 なお、選択手段340は、予測精度が最も高い一の推奨組み合わせ情報を選択する代わりに、予測精度が予め定められた基準よりも高い推奨組み合わせ情報を選択してもよい。また、選択手段340は、評価結果が最も優れる一の推奨組み合わせ情報を選択する代わりに、評価結果が予め定められた基準よりも優れる推奨組み合わせ情報を選択してもよい。これらの場合、出力手段370は、選択された全ての推奨組み合わせ情報を出力する。その際、予測精度が最も高い一の推奨組み合わせ情報または評価結果が最も優れる一の推奨組み合わせ情報が他の推奨組み合わせ情報と識別可能な状態で出力されても良い。 Note that the selection unit 340 may select recommended combination information whose prediction accuracy is higher than a predetermined reference instead of selecting one recommended combination information having the highest prediction accuracy. Further, instead of selecting one recommended combination information having the best evaluation result, the selection means 340 may select recommended combination information having the evaluation result better than a predetermined reference. In these cases, the output unit 370 outputs all the selected recommended combination information. At this time, one recommended combination information with the highest prediction accuracy or one recommended combination information with the best evaluation result may be output in a state that can be distinguished from other recommended combination information.
 また、処理装置30は選択手段340を備えなくても良い。この場合、出力手段370は抽出手段320に抽出された推奨条件に関連づけられた推奨組み合わせ情報を全て出力する。 (4) The processing device 30 may not include the selection unit 340. In this case, the output unit 370 outputs all the recommended combination information associated with the extracted recommended condition to the extraction unit 320.
 また、抽出手段320に抽出される推奨条件の数がゼロである場合、または、選択手段340で選択される推奨組み合わせ情報の数がゼロである場合、出力手段370は適当な組み合わせがないことを示す情報を出力する。 When the number of recommended conditions extracted by the extraction unit 320 is zero, or when the number of recommended combination information selected by the selection unit 340 is zero, the output unit 370 determines that there is no appropriate combination. Outputs the indicated information.
 本実施形態に係る処理装置30は、図7に示したような計算機1000により実現可能である。本実施形態において、ストレージデバイス1080は、処理装置30の各機能構成部を実現するプログラムモジュールを記憶している。処理装置30は、情報処理装置20を実現するために用いられる計算機と同じ計算機で実現されてもよいし、異なる計算機で実現されてもよい。 処理 The processing device 30 according to the present embodiment can be realized by a computer 1000 as shown in FIG. In the present embodiment, the storage device 1080 stores a program module that implements each functional component of the processing device 30. The processing device 30 may be realized by the same computer as that used to realize the information processing device 20, or may be realized by a different computer.
 次に、本実施形態の作用および効果について説明する。本実施形態によれば、推奨情報を用いて、特定の検出環境に対し適したセンサ10の組み合わせを知ることができる。ひいては、におい成分の検出および検出結果に基づく予測を精度の良く行うことができる。 Next, the operation and effect of the present embodiment will be described. According to the present embodiment, it is possible to know a combination of the sensors 10 suitable for a specific detection environment using the recommended information. Consequently, the detection of the odor component and the prediction based on the detection result can be performed with high accuracy.
 (第4の実施形態)
 図12は、第4の実施形態に係る処理装置40の構成を例示する図である。本実施形態に係る処理装置40は、センサ10の検出環境の推奨条件と一以上のセンサ10からなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、使用可能なセンサ10を示す情報とに基づいて、推奨条件を出力する。ここで、推奨情報は、複数種類のセンサ10の集合100からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である。
(Fourth embodiment)
FIG. 12 is a diagram illustrating a configuration of a processing device 40 according to the fourth embodiment. The processing device 40 according to the present embodiment includes recommended information in which the recommended conditions of the detection environment of the sensor 10 and recommended combination information indicating a combination of one or more sensors 10 are associated with each other, and information indicating the available sensors 10. Output the recommended conditions based on this. Here, the recommendation information is information based on an execution result of machine learning that inputs a plurality of feature amounts based on outputs from a set 100 of a plurality of types of sensors 10 and correct answer data.
 本図の例において、処理装置40は、抽出手段420、選択手段440および出力手段470を備える。抽出手段420は、推奨情報に含まれる推奨組み合わせ情報から、使用可能なセンサ10を示す情報に含まれるセンサ10で、実現可能な組み合わせを示す推奨組み合わせ情報を抽出する。選択手段440は、抽出された推奨組み合わせ情報に関連づけられた推奨条件から、一以上の推奨条件を選択する。そして、出力手段470は、選択された推奨条件を出力する。 処理 In the example of this drawing, the processing device 40 includes an extraction unit 420, a selection unit 440, and an output unit 470. The extracting unit 420 extracts, from the recommended combination information included in the recommended information, recommended combination information indicating a feasible combination with the sensor 10 included in the information indicating the available sensor 10. The selection unit 440 selects one or more recommended conditions from the recommended conditions associated with the extracted recommended combination information. Then, the output unit 470 outputs the selected recommended condition.
 図13は、第4の実施形態に係る処理方法を例示するフローチャートである。本処理方法では、センサの検出環境の推奨条件と一以上のセンサ10からなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、使用可能なセンサ10を示す情報とに基づいて、推奨条件が出力される。 FIG. 13 is a flowchart illustrating a processing method according to the fourth embodiment. In this processing method, the recommended condition is determined based on the recommended information that associates the recommended condition of the sensor detection environment with the recommended combination information indicating a combination of one or more sensors 10 and the information indicating the available sensors 10. Is output.
 本図の例において、本実施形態に係る処理方法は、抽出ステップS420、選択ステップS440、および出力ステップS470を含む。抽出ステップS420では、推奨情報に含まれる推奨組み合わせ情報から、使用可能なセンサ10を示す情報に含まれるセンサ10で、実現可能な組み合わせを示す推奨組み合わせ情報が抽出される。選択ステップS440では、抽出された推奨組み合わせ情報に関連づけられた推奨条件から、一以上の推奨条件が選択される。そして、出力ステップS470では、選択された推奨条件が出力される。 に お い て In the example of this drawing, the processing method according to the present embodiment includes an extraction step S420, a selection step S440, and an output step S470. In the extraction step S420, the recommended combination information indicating a feasible combination by the sensor 10 included in the information indicating the usable sensor 10 is extracted from the recommended combination information included in the recommended information. In the selection step S440, one or more recommended conditions are selected from the recommended conditions associated with the extracted recommended combination information. Then, in the output step S470, the selected recommended condition is output.
 本実施形態に係る推奨情報は、たとえば第1および第2の実施形態の少なくともいずれかに係る推奨情報生成手段270で生成される推奨情報と同じである。また、以下の説明において情報処理装置20は、第1および第2の実施形態の少なくともいずれかに係る情報処理装置20と同じである。本実施形態に係る処理装置40および処理方法によれば、推奨情報を用いて、特定のセンサ10の組み合わせを用いる際に好ましい検出環境の条件を求めることができる。以下に詳しく説明する。 The recommended information according to the present embodiment is the same as the recommended information generated by the recommended information generating unit 270 according to at least one of the first and second embodiments, for example. Further, in the following description, the information processing device 20 is the same as the information processing device 20 according to at least one of the first and second embodiments. According to the processing device 40 and the processing method according to the present embodiment, it is possible to use the recommendation information to obtain a preferable detection environment condition when a specific combination of sensors 10 is used. This will be described in detail below.
 抽出ステップS420において抽出手段420は、推奨情報および使用可能なセンサ10を示す情報を取得する。抽出手段420は推奨情報を、抽出手段420からアクセス可能な記憶装置または情報処理装置20から取得できる。また、処理装置40のユーザは処理装置40に対して使用可能なセンサ10を示す情報を入力可能である。そして、抽出手段420は入力された、使用可能なセンサ10を示す情報を取得する。使用可能なセンサ10を示す情報は、複数のセンサ10を示しても良い。使用可能なセンサ10を示す情報はたとえば、ユーザがセンサモジュール等で同時に使用可能なセンサ10の組み合わせを示す。使用可能なセンサ10を示す情報に複数のセンサ10が含まれる場合、これらのセンサ10の種類は互いに異なる。すなわち、これらのセンサ10の官能膜等が互いに異なる。 In the extraction step S420, the extraction means 420 acquires the recommended information and the information indicating the usable sensor 10. The extracting unit 420 can obtain the recommended information from a storage device or the information processing device 20 accessible from the extracting unit 420. Further, the user of the processing device 40 can input information indicating the sensors 10 that can be used for the processing device 40. Then, the extracting unit 420 acquires the input information indicating the usable sensor 10. The information indicating the usable sensors 10 may indicate a plurality of sensors 10. The information indicating the usable sensors 10 indicates, for example, a combination of the sensors 10 that can be simultaneously used by a user in a sensor module or the like. When the information indicating the usable sensors 10 includes a plurality of sensors 10, the types of these sensors 10 are different from each other. That is, the functional films and the like of these sensors 10 are different from each other.
 第1および第2の実施形態において説明した通り、推奨情報には一以上の推奨組み合わせ情報と、それに関連づけられた推奨条件が含まれる。抽出手段420は、推奨情報の中から、使用可能なセンサ10を示す情報で実現可能な組み合わせを示す推奨組み合わせ情報を全て抽出する。なお、抽出された推奨組み合わせ情報が示す組み合わせには、使用可能なセンサ10を示す情報に示されるセンサ10に加え、他のセンサ10がさらに含まれても良い。 As described in the first and second embodiments, the recommended information includes one or more recommended combination information and a recommended condition associated therewith. The extracting unit 420 extracts all pieces of recommended combination information indicating combinations that can be realized by the information indicating the available sensors 10 from the recommended information. The combination indicated by the extracted recommended combination information may further include another sensor 10 in addition to the sensor 10 indicated by the information indicating the usable sensor 10.
 次いで選択手段440は選択ステップS440において、抽出手段420で抽出された推奨組み合わせ情報に関連づけられた推奨条件のうち、特に有用性が高い推奨条件を選択する。 Next, in the selection step S440, the selection unit 440 selects a recommended condition having a particularly high usefulness from among the recommended conditions associated with the recommended combination information extracted by the extraction unit 420.
 たとえば選択手段440は、予測精度に基づいて、一以上の推奨条件を選択することができる。具体的には、推奨情報において、推奨条件および推奨組み合わせ情報には、その推奨条件とその推奨組み合わせ情報が示す組み合わせとを用いた場合の、におい成分に関する予測精度がさらに関連づけられていてもよい。すると、選択手段440は、抽出された推奨組み合わせ情報に関連づけられた推奨条件のうち、予測精度が最も高い推奨条件を選択することができる。 {For example, the selection unit 440 can select one or more recommended conditions based on the prediction accuracy. Specifically, in the recommended information, the recommended condition and the recommended combination information may be further associated with the prediction accuracy of the odor component when the recommended condition and the combination indicated by the recommended combination information are used. Then, the selection unit 440 can select a recommended condition having the highest prediction accuracy from among the recommended conditions associated with the extracted recommended combination information.
 また、選択手段440はたとえば、推奨条件の広さおよび予め定められた条件(標準値)からの近さの少なくとも一方に基づいて、一以上の推奨条件を選択してもよい。具体的には、推奨情報において、推奨条件および推奨組み合わせ情報に、推奨条件の広さおよび予め定められた条件からの近さの少なくとも一方に基づく評価結果がさらに関連づけられていても良い。すると、選択手段440は、抽出された推奨組み合わせ情報に関連づけられた推奨条件のうち、評価結果が最も優れる推奨条件を選択することができる。 The selection unit 440 may select one or more recommended conditions based on, for example, at least one of the size of the recommended condition and the proximity to a predetermined condition (standard value). Specifically, in the recommended information, an evaluation result based on at least one of the size of the recommended condition and the proximity to a predetermined condition may be further associated with the recommended condition and the recommended combination information. Then, the selecting unit 440 can select the recommended condition having the best evaluation result among the recommended conditions associated with the extracted recommended combination information.
 なお、推奨条件の広さおよび予め定められた条件からの近さの少なくとも一方に基づく評価結果は選択手段440で算出されても良い。この場合、選択手段440は、情報処理装置20の評価手段240が評価ステップS240において行ったのと同様の方法で評価結果を算出することができる。そして、選択手段440は算出した評価結果に基づいて推奨条件を選択する。 The evaluation result based on at least one of the size of the recommended condition and the proximity to the predetermined condition may be calculated by the selection unit 440. In this case, the selection unit 440 can calculate the evaluation result by the same method as that performed by the evaluation unit 240 of the information processing device 20 in the evaluation step S240. Then, the selection unit 440 selects a recommended condition based on the calculated evaluation result.
 次いで、出力ステップS470において出力手段470は、選択手段440に選択された推奨条件を出力する。出力手段470はたとえば、推奨条件を出力手段470からアクセス可能な記憶装置に記憶させても良いし、外部の装置に対して出力しても良いし、処理装置40に接続された表示装置に表示させても良い。処理装置40のユーザは、出力された推奨条件に基づき、測定の際の検出環境を決定することができる。そしてユーザはたとえば、決定した検出環境を実現するよう、測定において温度や湿度等を調整する。 Next, in the output step S470, the output unit 470 outputs the selected recommended condition to the selection unit 440. The output unit 470 may, for example, store the recommended conditions in a storage device accessible from the output unit 470, output the recommended conditions to an external device, or display the recommended conditions on a display device connected to the processing device 40. You may let it. The user of the processing device 40 can determine the detection environment at the time of measurement based on the output recommended conditions. Then, for example, the user adjusts temperature, humidity, and the like in the measurement so as to realize the determined detection environment.
 また、出力手段470は、推奨条件に関連づけて、予測式を示す情報をさらに出力しても良い。予測式を示す情報はたとえば推奨条件に関連づけられた状態で推奨情報に含まれる。ユーザは、検出環境に応じて、出力された予測式を用い、におい成分に関する予測を行うことができる。 (4) The output unit 470 may further output information indicating a prediction formula in association with the recommended condition. Information indicating the prediction formula is included in the recommended information, for example, in a state associated with the recommended condition. The user can predict the odor component using the output prediction formula according to the detection environment.
 なお、選択手段440は、予測精度が最も高い一の推奨条件を選択する代わりに、予測精度が予め定められた基準よりも高い推奨条件を選択してもよい。また、選択手段440は、評価結果が最も優れる一の推奨条件を選択する代わりに、評価結果が予め定められた基準よりも優れる推奨条件を選択してもよい。これらの場合、出力手段470は、選択された全ての推奨条件を出力する。その際、予測精度が最も高い一の推奨条件または評価結果が最も優れる一の推奨条件が他の推奨条件と識別可能な状態で出力されても良い。 Note that, instead of selecting one recommended condition having the highest prediction accuracy, the selection unit 440 may select a recommended condition whose prediction accuracy is higher than a predetermined reference. In addition, instead of selecting one recommended condition having the best evaluation result, the selection unit 440 may select a recommended condition whose evaluation result is better than a predetermined reference. In these cases, the output unit 470 outputs all the selected recommended conditions. At this time, one recommended condition with the highest prediction accuracy or one recommended condition with the best evaluation result may be output in a state that can be distinguished from other recommended conditions.
 また、処理装置40は選択手段440を備えなくても良い。この場合、出力手段470は抽出手段420に抽出された推奨組み合わせ情報に関連づけられた推奨条件を全て出力する。 (4) The processing device 40 may not include the selection unit 440. In this case, the output unit 470 outputs all the recommended conditions associated with the extracted recommended combination information to the extraction unit 420.
 また、抽出手段420に抽出される推奨組み合わせ情報の数がゼロである場合、または、選択手段440で選択される推奨条件の数がゼロである場合、出力手段470は適当な条件がないことを示す情報を出力する。 When the number of recommended combination information extracted by the extraction unit 420 is zero, or when the number of recommended conditions selected by the selection unit 440 is zero, the output unit 470 determines that there is no appropriate condition. Outputs the indicated information.
 本実施形態に係る処理装置40は、図7に示したような計算機1000により実現可能である。本実施形態において、ストレージデバイス1080は、処理装置40の各機能構成部を実現するプログラムモジュールを記憶している。処理装置40は、情報処理装置20を実現するために用いられる計算機と同じ計算機で実現されてもよいし、異なる計算機で実現されてもよい。 The processing device 40 according to the present embodiment can be realized by a computer 1000 as shown in FIG. In the present embodiment, the storage device 1080 stores a program module that implements each functional component of the processing device 40. The processing device 40 may be realized by the same computer as the computer used to realize the information processing device 20, or may be realized by a different computer.
 次に、本実施形態の作用および効果について説明する。本実施形態によれば、推奨情報を用いて、使用するセンサ10に対し好ましい検出環境の条件を知ることができる。ひいては、その条件に従って、におい成分の検出および検出結果に基づく予測を精度の良く行うことができる。 Next, the operation and effect of the present embodiment will be described. According to the present embodiment, it is possible to know the conditions of the detection environment preferable for the sensor 10 to be used, using the recommended information. As a result, detection of the odor component and prediction based on the detection result can be performed with high accuracy according to the condition.
 以上、図面を参照して本発明の実施形態について述べたが、これらは本発明の例示であり、上記以外の様々な構成を採用することもできる。たとえば、上述の説明で用いたシーケンス図やフローチャートでは、複数の工程(処理)が順番に記載されているが、各実施形態で実行される工程の実行順序は、その記載の順番に制限されない。各実施形態では、図示される工程の順番を内容的に支障のない範囲で変更することができる。また、上述の各実施形態は、内容が相反しない範囲で組み合わせることができる。 Although the embodiments of the present invention have been described above with reference to the drawings, they are merely examples of the present invention, and various configurations other than the above can be adopted. For example, in the sequence diagrams and flowcharts used in the above description, a plurality of steps (processes) are described in order, but the execution order of the steps executed in each embodiment is not limited to the described order. In each embodiment, the order of the illustrated steps can be changed within a range that does not hinder the contents. In addition, the above-described embodiments can be combined in a range where the contents do not conflict with each other.
 以上、図面を参照して本発明の実施形態について述べたが、これらは本発明の例示であり、上記以外の様々な構成を採用することもできる。 Although the embodiments of the present invention have been described above with reference to the drawings, they are merely examples of the present invention, and various configurations other than the above can be adopted.
 上記の実施形態の一部又は全部は、以下の付記のようにも記載されうるが、以下には限られない。
1. 複数種類のセンサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づき、前記センサの検出環境の推奨条件と、一以上の前記センサからなる組み合わせとを関連づけた推奨情報を生成する推奨情報生成手段を備える情報処理装置。
2. 1.に記載の情報処理装置において、
 前記機械学習を行うことにより、前記複数の特徴量を変数とする式であって、におい成分に関する予測を行うための予測式を生成する予測式生成手段と、
 前記予測式における前記複数の特徴量に対する複数の重みに基づいて、前記集合から一以上の前記センサを抽出し、抽出した前記センサからなる前記組み合わせを示す推奨組み合わせ情報を生成する抽出手段とをさらに備え、
 前記抽出手段は、前記予測式において、前記複数の重みのうち予め定められた条件を満たす、または満たさない前記重みで重みづけられた前記特徴量の、出力元である前記センサを抽出し、
 前記推奨情報は、一以上の前記推奨組み合わせ情報を含む情報処理装置。
3. 2.に記載の情報処理装置において、
 前記予測式生成手段は、
  前記検出環境に基づいた分岐を含むモデルを用いて前記予測式を生成し、
  前記推奨組み合わせ情報に、前記予測式に適した前記検出環境の条件であって、前記分岐の条件に基づく前記検出環境の条件を前記推奨条件として関連づける情報処理装置。
4. 3.に記載の情報処理装置において、
 前記機械学習は、前記特徴量に関連づけられた前記検出環境をさらに入力とした異種混合学習であり、
 前記分岐の条件は、前記異種混合学習により生成される情報処理装置。
5. 2.から4.のいずれか一つに記載の情報処理装置において、
 前記検出環境は、温度、湿度、気圧、夾雑ガスの種類、パージガスの種類、サンプリング周期、対象物と前記センサとの距離、前記センサの周囲に存在する物体のうち少なくともいずれかを含む情報処理装置。
6. 2.から5.のいずれか一つに記載の情報処理装置において、
 前記予測式の予測精度を算出する予測精度算出手段をさらに備え、
 前記推奨情報は、前記推奨組み合わせ情報に関連づけられた前記予測式の前記予測精度をさらに含む情報処理装置。
7. 2.から6.のいずれか一つに記載の情報処理装置において、
 前記組み合わせを採用する場合のコストに少なくとも基づいて、前記組み合わせを評価する評価手段をさらに備え、
 前記推奨情報は、前記推奨組み合わせ情報に関連づけられた、前記評価手段の評価結果をさらに含む情報処理装置。
8. 7.に記載の情報処理装置において、
 前記評価手段は、前記推奨組み合わせ情報に関連づけられた前記推奨条件にさらに基づいて、前記組み合わせを評価する情報処理装置。
9. 1.から8.のいずれか一つに記載の情報処理装置で生成された前記推奨情報と、前記検出環境を示す情報とに基づいて、使用する前記センサを決定する決定方法。
10. 1.から8.のいずれか一つに記載の情報処理装置で生成された前記推奨情報と、使用可能な前記センサを示す情報とに基づいて、前記検出環境を決定する決定方法。
11. センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、前記検出環境を示す情報とに基づいて、前記組み合わせを出力し、
 前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である処理装置。
12. 11.に記載の処理装置において、
 前記推奨情報に含まれる前記推奨条件から、前記検出環境を示す情報が適合する前記推奨条件を抽出する抽出手段と、
 抽出された前記推奨条件に関連づけられた前記推奨組み合わせ情報から、一以上の前記推奨組み合わせ情報を選択する選択手段と、
 選択された前記推奨組み合わせ情報が示す前記組み合わせを出力する出力手段とを備える処理装置。
13. 12.に記載の処理装置において、
 前記推奨情報において、前記推奨条件および前記推奨組み合わせ情報には、当該推奨条件と当該推奨組み合わせ情報が示す前記組み合わせとを用いた場合の、におい成分に関する予測精度がさらに関連づけられており、
 前記選択手段は、抽出された前記推奨条件に関連づけられた前記推奨組み合わせ情報から、前記予測精度に基づいて、一以上の前記推奨組み合わせ情報を選択する処理装置。
14. 12.または13.に記載の処理装置において、
 前記選択手段は、前記推奨組み合わせ情報が示す前記組み合わせを用いた場合の、コストに基づいて、一以上の前記推奨組み合わせ情報を選択する処理装置。
15. センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、使用可能な前記センサを示す情報とに基づいて、前記推奨条件を出力し、
 前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である処理装置。
16. 15.に記載の処理装置において、
 前記推奨情報に含まれる前記推奨組み合わせ情報から、前記使用可能な前記センサを示す情報に含まれる前記センサで、実現可能な前記組み合わせを示す前記推奨組み合わせ情報を抽出する抽出手段と、
 抽出された前記推奨組み合わせ情報に関連づけられた前記推奨条件から、一以上の前記推奨条件を選択する選択手段と、
 選択された前記推奨条件を出力する出力手段とを備える処理装置。
17. 16.に記載の処理装置において、
 前記推奨情報において、前記推奨条件および前記推奨組み合わせ情報には、当該推奨条件と当該推奨組み合わせ情報が示す前記組み合わせとを用いた場合の、におい成分に関する予測精度がさらに関連づけられており、
 前記選択手段は、前記予測精度に基づいて、一以上の前記推奨条件を選択する処理装置。
18. 16.または17.に記載の処理装置において、
 前記選択手段は、前記推奨条件の広さおよび予め定められた条件からの近さの少なくとも一方に基づいて、一以上の前記推奨条件を選択する処理装置。
19. 11.から18.のいずれか一つに記載の処理装置において、
 前記検出環境は、温度、湿度、気圧、夾雑ガスの種類、パージガスの種類、サンプリング周期、対象物と前記センサとの距離、前記センサの周囲に存在する物体のうち少なくともいずれかを含む処理装置。
20. 複数種類のセンサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づき、前記センサの検出環境の推奨条件と、一以上の前記センサからなる組み合わせとを関連づけた推奨情報を生成する推奨情報生成ステップを含む情報処理方法。
21. 20.に記載の情報処理方法において、
 前記機械学習を行うことにより、前記複数の特徴量を変数とする式であって、におい成分に関する予測を行うための予測式を生成する予測式生成ステップと、
 前記予測式における前記複数の特徴量に対する複数の重みに基づいて、前記集合から一以上の前記センサを抽出し、抽出した前記センサからなる前記組み合わせを示す推奨組み合わせ情報を生成する抽出ステップとをさらに含み、
 前記抽出ステップでは、前記予測式において、前記複数の重みのうち予め定められた条件を満たす、または満たさない前記重みで重みづけられた前記特徴量の、出力元である前記センサを抽出し、
 前記推奨情報は、一以上の前記推奨組み合わせ情報を含む情報処理方法。
22. 21.に記載の情報処理方法において、
 前記予測式生成ステップでは、
  前記検出環境に基づいた分岐を含むモデルを用いて前記予測式を生成し、
  前記推奨組み合わせ情報に、前記予測式に適した前記検出環境の条件であって、前記分岐の条件に基づく前記検出環境の条件を前記推奨条件として関連づける情報処理方法。
23. 22.に記載の情報処理方法において、
 前記機械学習は、前記特徴量に関連づけられた前記検出環境をさらに入力とした異種混合学習であり、
 前記分岐の条件は、前記異種混合学習により生成される情報処理方法。
24. 21.から23.のいずれか一つに記載の情報処理方法において、
 前記検出環境は、温度、湿度、気圧、夾雑ガスの種類、パージガスの種類、サンプリング周期、対象物と前記センサとの距離、前記センサの周囲に存在する物体のうち少なくともいずれかを含む情報処理方法。
25. 21.から24.のいずれか一つに記載の情報処理方法において、
 前記予測式の予測精度を算出する予測精度算出ステップをさらに含み、
 前記推奨情報は、前記推奨組み合わせ情報に関連づけられた前記予測式の前記予測精度をさらに含む情報処理方法。
26. 21.から25.のいずれか一つに記載の情報処理方法において、
 前記組み合わせを採用する場合のコストに少なくとも基づいて、前記組み合わせを評価する評価ステップをさらに含み、
 前記推奨情報は、前記推奨組み合わせ情報に関連づけられた、前記評価ステップでの評価結果をさらに含む情報処理方法。
27. 26.に記載の情報処理方法において、
 前記評価ステップでは、前記推奨組み合わせ情報に関連づけられた前記推奨条件にさらに基づいて、前記組み合わせを評価する情報処理方法。
28. センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、前記検出環境を示す情報とに基づいて、前記組み合わせを出力し、
 前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である処理方法。
29. 28.に記載の処理方法において、
 前記推奨情報に含まれる前記推奨条件から、前記検出環境を示す情報が適合する前記推奨条件を抽出する抽出ステップと、
 抽出された前記推奨条件に関連づけられた前記推奨組み合わせ情報から、一以上の前記推奨組み合わせ情報を選択する選択ステップと、
 選択された前記推奨組み合わせ情報が示す前記組み合わせを出力する出力ステップとを含む処理方法。
30. 29.に記載の処理方法において、
 前記推奨情報において、前記推奨条件および前記推奨組み合わせ情報には、当該推奨条件と当該推奨組み合わせ情報が示す前記組み合わせとを用いた場合の、におい成分に関する予測精度がさらに関連づけられており、
 前記選択ステップでは、抽出された前記推奨条件に関連づけられた前記推奨組み合わせ情報から、前記予測精度に基づいて、一以上の前記推奨組み合わせ情報を選択する処理方法。
31. 29.または30.に記載の処理方法において、
 前記選択ステップでは、前記推奨組み合わせ情報が示す前記組み合わせを用いた場合の、コストに基づいて、一以上の前記推奨組み合わせ情報を選択する処理方法。
32. センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、使用可能な前記センサを示す情報とに基づいて、前記推奨条件を出力し、
 前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である処理方法。
33. 32.に記載の処理方法において、
 前記推奨情報に含まれる前記推奨組み合わせ情報から、前記使用可能な前記センサを示す情報に含まれる前記センサで、実現可能な前記組み合わせを示す前記推奨組み合わせ情報を抽出する抽出ステップと、
 抽出された前記推奨組み合わせ情報に関連づけられた前記推奨条件から、一以上の前記推奨条件を選択する選択ステップと、
 選択された前記推奨条件を出力する出力ステップとを含む処理方法。
34. 33.に記載の処理方法において、
 前記推奨情報において、前記推奨条件および前記推奨組み合わせ情報には、当該推奨条件と当該推奨組み合わせ情報が示す前記組み合わせとを用いた場合の、におい成分に関する予測精度がさらに関連づけられており、
 前記選択ステップでは、前記予測精度に基づいて、一以上の前記推奨条件を選択する処理方法。
35. 33.または34.に記載の処理方法において、
 前記選択ステップでは、前記推奨条件の広さおよび予め定められた条件からの近さの少なくとも一方に基づいて、一以上の前記推奨条件を選択する処理方法。
36. 28.から35.のいずれか一つに記載の処理方法において、
 前記検出環境は、温度、湿度、気圧、夾雑ガスの種類、パージガスの種類、サンプリング周期、対象物と前記センサとの距離、前記センサの周囲に存在する物体のうち少なくともいずれかを含む処理方法。
37. 20.から27.のいずれか一つに記載の情報処理方法の各ステップをコンピュータに実行させるプログラム。
38. 28.から36.のいずれか一つに記載の処理方法の各ステップをコンピュータに実行させるプログラム。
Some or all of the above embodiments may be described as in the following supplementary notes, but are not limited thereto.
1. Based on execution results of machine learning using a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data, based on a recommended condition of a detection environment of the sensor and a combination of one or more of the sensors An information processing apparatus including a recommended information generating unit that generates recommended information in which the information is associated with the recommended information.
2. 1. In the information processing device according to the above,
By performing the machine learning, a prediction formula generating means for generating a formula using the plurality of feature amounts as variables, and a prediction formula for performing prediction regarding an odor component,
Extracting means for extracting one or more of the sensors from the set based on a plurality of weights for the plurality of feature amounts in the prediction formula, and generating recommended combination information indicating the combination of the extracted sensors. Prepared,
The extracting means, in the prediction formula, satisfying a predetermined condition among the plurality of weights, or extracting the sensor as an output source of the feature amount weighted by the weight that does not satisfy,
The information processing device, wherein the recommended information includes one or more pieces of the recommended combination information.
3. 2. In the information processing device according to the above,
The prediction formula generation means includes:
Using the model including a branch based on the detection environment to generate the prediction formula,
An information processing apparatus for associating the recommended combination information with a condition of the detection environment suitable for the prediction formula, the condition of the detection environment based on the condition of the branch, as the recommended condition.
4. 3. In the information processing device according to the above,
The machine learning is a heterogeneous mixture learning further inputting the detection environment associated with the feature amount,
The information processing device according to claim 1, wherein the branch condition is generated by the heterogeneous learning.
5. 2. From 4. In the information processing apparatus according to any one of the above,
The information processing device includes at least one of a temperature, a humidity, an atmospheric pressure, a type of a contaminant gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor. .
6. 2. To 5. In the information processing apparatus according to any one of the above,
The apparatus further includes a prediction accuracy calculation unit that calculates prediction accuracy of the prediction expression,
The information processing apparatus, wherein the recommendation information further includes the prediction accuracy of the prediction formula associated with the recommendation combination information.
7. 2. From 6. In the information processing apparatus according to any one of the above,
At least based on the cost of employing the combination, further comprising an evaluation means for evaluating the combination,
The information processing device, wherein the recommendation information further includes an evaluation result of the evaluation unit, which is associated with the recommendation combination information.
8. 7. In the information processing device according to the above,
The information processing device, wherein the evaluation unit evaluates the combination based on the recommended condition associated with the recommended combination information.
9. 1. From 8. A determination method for determining the sensor to be used based on the recommendation information generated by the information processing apparatus according to any one of the above and information indicating the detection environment.
10. 1. From 8. A determination method for determining the detection environment based on the recommendation information generated by the information processing apparatus according to any one of the above and information indicating the usable sensor.
11. Based on information indicating the detection environment and the recommended information relating the recommended conditions of the sensor detection environment and the recommended combination information indicating the combination of one or more sensors, and outputting the combination,
The processing device, wherein the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data.
12. 11. In the processing apparatus described in the above,
From the recommended conditions included in the recommended information, extraction means for extracting the recommended conditions to which the information indicating the detection environment is suitable,
From the recommended combination information associated with the extracted recommended conditions, selecting means for selecting one or more of the recommended combination information,
Output means for outputting the combination indicated by the selected recommended combination information.
13. 12. In the processing apparatus described in the above,
In the recommended information, the recommended conditions and the recommended combination information, when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated,
The processing device, wherein the selection unit selects one or more pieces of the recommended combination information based on the prediction accuracy from the extracted recommended combination information associated with the extracted recommended condition.
14. 12. Or 13. In the processing apparatus described in the above,
The processing device, wherein the selection unit selects one or more pieces of the recommended combination information based on a cost when the combination indicated by the recommended combination information is used.
15. Based on information indicating the recommended conditions of the detection environment of the sensor and recommended combination information indicating a combination of one or more sensors, and information indicating the available sensors, the recommended conditions are output.
The processing device, wherein the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data.
16. 15. In the processing apparatus described in the above,
From the recommended combination information included in the recommended information, the sensor included in the information indicating the available sensor, the extraction unit that extracts the recommended combination information indicating the feasible combination,
From the recommended conditions associated with the extracted recommended combination information, selecting means for selecting one or more of the recommended conditions,
Output means for outputting the selected recommended condition.
17. 16. In the processing apparatus described in the above,
In the recommended information, the recommended conditions and the recommended combination information, when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated,
The processing device, wherein the selecting unit selects one or more of the recommended conditions based on the prediction accuracy.
18. 16. Or 17. In the processing apparatus described in the above,
The processing device, wherein the selecting unit selects one or more of the recommended conditions based on at least one of a size of the recommended condition and a proximity to a predetermined condition.
19. 11. From 18. In the processing apparatus according to any one of the above,
The processing apparatus according to claim 1, wherein the detection environment includes at least one of a temperature, a humidity, an atmospheric pressure, a type of an impurity gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor.
20. Based on execution results of machine learning using a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data, based on a recommended condition of a detection environment of the sensor and a combination of one or more of the sensors An information processing method that includes a recommended information generating step of generating recommended information in which recommended information is associated with.
21. 20. In the information processing method described in the above,
By performing the machine learning, a prediction formula generating step of generating a prediction formula for performing a prediction on an odor component, wherein the prediction formula generation step is a formula using the plurality of feature amounts as variables,
Extracting one or more of the sensors from the set based on a plurality of weights for the plurality of feature amounts in the prediction formula, and generating recommended combination information indicating the combination of the extracted sensors. Including
In the extracting step, in the prediction equation, satisfying a predetermined condition among the plurality of weights, or extracting the sensor as an output source of the feature amount weighted by the weight that does not satisfy,
The information processing method, wherein the recommended information includes one or more pieces of the recommended combination information.
22. 21. In the information processing method described in the above,
In the prediction formula generation step,
Using the model including a branch based on the detection environment to generate the prediction formula,
An information processing method for associating the recommended combination information with a condition of the detection environment suitable for the prediction equation, the condition of the detection environment based on the condition of the branch, as the recommended condition.
23. 22. In the information processing method described in the above,
The machine learning is a heterogeneous mixture learning further inputting the detection environment associated with the feature amount,
The information processing method according to claim 1, wherein the branch condition is generated by the heterogeneous mixture learning.
24. 21. To 23. In the information processing method according to any one of the above,
The information processing method includes at least one of a temperature, a humidity, an atmospheric pressure, a type of an impurity gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor. .
25. 21. To 24. In the information processing method according to any one of the above,
The method further includes a prediction accuracy calculation step of calculating the prediction accuracy of the prediction expression,
The information processing method, wherein the recommendation information further includes the prediction accuracy of the prediction formula associated with the recommendation combination information.
26. 21. To 25. In the information processing method according to any one of the above,
At least based on the cost of employing the combination, further comprising an evaluation step of evaluating the combination,
The information processing method, wherein the recommendation information further includes an evaluation result in the evaluation step, which is associated with the recommendation combination information.
27. 26. In the information processing method described in the above,
In the evaluating step, an information processing method for evaluating the combination based on the recommended condition associated with the recommended combination information.
28. Based on the recommended conditions of the sensor detection environment and the recommended information associated with the recommended combination information indicating a combination of one or more sensors, and the information indicating the detection environment, the combination is output.
A processing method, wherein the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data.
29. 28. In the processing method described in
From the recommended conditions included in the recommended information, an extraction step of extracting the recommended conditions to which information indicating the detection environment is suitable,
A selection step of selecting one or more of the recommended combination information from the recommended combination information associated with the extracted recommended condition,
Outputting the combination indicated by the selected recommended combination information.
30. 29. In the processing method described in
In the recommended information, the recommended conditions and the recommended combination information, when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated,
In the selecting step, a processing method of selecting one or more pieces of the recommended combination information from the recommended combination information associated with the extracted recommended condition based on the prediction accuracy.
31. 29. Or 30. In the processing method described in
In the selecting step, a processing method for selecting one or more pieces of the recommended combination information based on a cost when the combination indicated by the recommended combination information is used.
32. Based on information indicating the recommended conditions of the detection environment of the sensor and recommended combination information indicating a combination of one or more sensors, and information indicating the available sensors, the recommended conditions are output.
A processing method, wherein the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data.
33. 32. In the processing method described in
From the recommended combination information included in the recommended information, the sensor included in the information indicating the available sensor, an extraction step of extracting the recommended combination information indicating the feasible combination,
A selecting step of selecting one or more of the recommended conditions from the recommended conditions associated with the extracted recommended combination information,
Outputting the selected recommended condition.
34. 33. In the processing method described in
In the recommended information, the recommended conditions and the recommended combination information, when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated,
In the selecting step, a processing method of selecting one or more recommended conditions based on the prediction accuracy.
35. 33. Or 34. In the processing method described in
In the selecting step, a processing method of selecting one or more of the recommended conditions based on at least one of a size of the recommended condition and a proximity from a predetermined condition.
36. 28. To 35. In the processing method according to any one of the above,
The processing method, wherein the detection environment includes at least one of a temperature, a humidity, an atmospheric pressure, a type of an impurity gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor.
37. 20. From 27. A program for causing a computer to execute each step of the information processing method according to any one of the above.
38. 28. From 36. A program for causing a computer to execute each step of the processing method according to any one of the above.

Claims (38)

  1.  複数種類のセンサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づき、前記センサの検出環境の推奨条件と、一以上の前記センサからなる組み合わせとを関連づけた推奨情報を生成する推奨情報生成手段を備える情報処理装置。 Based on execution results of machine learning using a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data, based on a recommended condition of a detection environment of the sensor and a combination of one or more of the sensors An information processing apparatus including a recommended information generating unit that generates recommended information in which the information is associated with the recommended information.
  2.  請求項1に記載の情報処理装置において、
     前記機械学習を行うことにより、前記複数の特徴量を変数とする式であって、におい成分に関する予測を行うための予測式を生成する予測式生成手段と、
     前記予測式における前記複数の特徴量に対する複数の重みに基づいて、前記集合から一以上の前記センサを抽出し、抽出した前記センサからなる前記組み合わせを示す推奨組み合わせ情報を生成する抽出手段とをさらに備え、
     前記抽出手段は、前記予測式において、前記複数の重みのうち予め定められた条件を満たす、または満たさない前記重みで重みづけられた前記特徴量の、出力元である前記センサを抽出し、
     前記推奨情報は、一以上の前記推奨組み合わせ情報を含む情報処理装置。
    The information processing apparatus according to claim 1,
    By performing the machine learning, a prediction formula generating means for generating a formula using the plurality of feature amounts as variables, and a prediction formula for performing prediction regarding an odor component,
    Extracting means for extracting one or more of the sensors from the set based on a plurality of weights for the plurality of feature amounts in the prediction formula, and generating recommended combination information indicating the combination of the extracted sensors. Prepared,
    The extracting means, in the prediction formula, satisfying a predetermined condition among the plurality of weights, or extracting the sensor as an output source of the feature amount weighted by the weight that does not satisfy,
    The information processing device, wherein the recommended information includes one or more pieces of the recommended combination information.
  3.  請求項2に記載の情報処理装置において、
     前記予測式生成手段は、
      前記検出環境に基づいた分岐を含むモデルを用いて前記予測式を生成し、
      前記推奨組み合わせ情報に、前記予測式に適した前記検出環境の条件であって、前記分岐の条件に基づく前記検出環境の条件を前記推奨条件として関連づける情報処理装置。
    The information processing apparatus according to claim 2,
    The prediction formula generation means includes:
    Using the model including a branch based on the detection environment to generate the prediction formula,
    An information processing apparatus for associating the recommended combination information with a condition of the detection environment suitable for the prediction formula, the condition of the detection environment based on the condition of the branch, as the recommended condition.
  4.  請求項3に記載の情報処理装置において、
     前記機械学習は、前記特徴量に関連づけられた前記検出環境をさらに入力とした異種混合学習であり、
     前記分岐の条件は、前記異種混合学習により生成される情報処理装置。
    The information processing device according to claim 3,
    The machine learning is a heterogeneous mixture learning further inputting the detection environment associated with the feature amount,
    The information processing device according to claim 1, wherein the branch condition is generated by the heterogeneous learning.
  5.  請求項2から4のいずれか一項に記載の情報処理装置において、
     前記検出環境は、温度、湿度、気圧、夾雑ガスの種類、パージガスの種類、サンプリング周期、対象物と前記センサとの距離、前記センサの周囲に存在する物体のうち少なくともいずれかを含む情報処理装置。
    The information processing apparatus according to any one of claims 2 to 4,
    The information processing device includes at least one of a temperature, a humidity, an atmospheric pressure, a type of a contaminant gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor. .
  6.  請求項2から5のいずれか一項に記載の情報処理装置において、
     前記予測式の予測精度を算出する予測精度算出手段をさらに備え、
     前記推奨情報は、前記推奨組み合わせ情報に関連づけられた前記予測式の前記予測精度をさらに含む情報処理装置。
    The information processing apparatus according to any one of claims 2 to 5,
    The apparatus further includes a prediction accuracy calculation unit that calculates prediction accuracy of the prediction expression,
    The information processing apparatus, wherein the recommendation information further includes the prediction accuracy of the prediction formula associated with the recommendation combination information.
  7.  請求項2から6のいずれか一項に記載の情報処理装置において、
     前記組み合わせを採用する場合のコストに少なくとも基づいて、前記組み合わせを評価する評価手段をさらに備え、
     前記推奨情報は、前記推奨組み合わせ情報に関連づけられた、前記評価手段の評価結果をさらに含む情報処理装置。
    The information processing apparatus according to any one of claims 2 to 6,
    At least based on the cost of employing the combination, further comprising an evaluation means for evaluating the combination,
    The information processing device, wherein the recommendation information further includes an evaluation result of the evaluation unit, which is associated with the recommendation combination information.
  8.  請求項7に記載の情報処理装置において、
     前記評価手段は、前記推奨組み合わせ情報に関連づけられた前記推奨条件にさらに基づいて、前記組み合わせを評価する情報処理装置。
    The information processing apparatus according to claim 7,
    The information processing device, wherein the evaluation unit evaluates the combination based on the recommended condition associated with the recommended combination information.
  9.  請求項1から8のいずれか一項に記載の情報処理装置で生成された前記推奨情報と、前記検出環境を示す情報とに基づいて、使用する前記センサを決定する決定方法。 A determination method for determining the sensor to be used based on the recommendation information generated by the information processing device according to any one of claims 1 to 8 and information indicating the detection environment.
  10.  請求項1から8のいずれか一項に記載の情報処理装置で生成された前記推奨情報と、使用可能な前記センサを示す情報とに基づいて、前記検出環境を決定する決定方法。 A determination method for determining the detection environment based on the recommendation information generated by the information processing apparatus according to any one of claims 1 to 8 and information indicating the available sensors.
  11.  センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、前記検出環境を示す情報とに基づいて、前記組み合わせを出力し、
     前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である処理装置。
    Based on information indicating the detection environment and the recommended information relating the recommended conditions of the sensor detection environment and the recommended combination information indicating the combination of one or more sensors, and outputting the combination,
    The processing device, wherein the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data.
  12.  請求項11に記載の処理装置において、
     前記推奨情報に含まれる前記推奨条件から、前記検出環境を示す情報が適合する前記推奨条件を抽出する抽出手段と、
     抽出された前記推奨条件に関連づけられた前記推奨組み合わせ情報から、一以上の前記推奨組み合わせ情報を選択する選択手段と、
     選択された前記推奨組み合わせ情報が示す前記組み合わせを出力する出力手段とを備える処理装置。
    The processing device according to claim 11,
    From the recommended conditions included in the recommended information, extraction means for extracting the recommended conditions to which the information indicating the detection environment is suitable,
    From the recommended combination information associated with the extracted recommended conditions, selecting means for selecting one or more of the recommended combination information,
    Output means for outputting the combination indicated by the selected recommended combination information.
  13.  請求項12に記載の処理装置において、
     前記推奨情報において、前記推奨条件および前記推奨組み合わせ情報には、当該推奨条件と当該推奨組み合わせ情報が示す前記組み合わせとを用いた場合の、におい成分に関する予測精度がさらに関連づけられており、
     前記選択手段は、抽出された前記推奨条件に関連づけられた前記推奨組み合わせ情報から、前記予測精度に基づいて、一以上の前記推奨組み合わせ情報を選択する処理装置。
    The processing device according to claim 12,
    In the recommended information, the recommended conditions and the recommended combination information, when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated,
    The processing device, wherein the selection unit selects one or more pieces of the recommended combination information based on the prediction accuracy from the extracted recommended combination information associated with the extracted recommended condition.
  14.  請求項12または13に記載の処理装置において、
     前記選択手段は、前記推奨組み合わせ情報が示す前記組み合わせを用いた場合の、コストに基づいて、一以上の前記推奨組み合わせ情報を選択する処理装置。
    The processing device according to claim 12 or 13,
    The processing device, wherein the selection unit selects one or more pieces of the recommended combination information based on a cost when the combination indicated by the recommended combination information is used.
  15.  センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、使用可能な前記センサを示す情報とに基づいて、前記推奨条件を出力し、
     前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である処理装置。
    Based on information indicating the recommended conditions of the detection environment of the sensor and recommended combination information indicating a combination of one or more sensors, and information indicating the available sensors, the recommended conditions are output.
    The processing device, wherein the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data.
  16.  請求項15に記載の処理装置において、
     前記推奨情報に含まれる前記推奨組み合わせ情報から、前記使用可能な前記センサを示す情報に含まれる前記センサで、実現可能な前記組み合わせを示す前記推奨組み合わせ情報を抽出する抽出手段と、
     抽出された前記推奨組み合わせ情報に関連づけられた前記推奨条件から、一以上の前記推奨条件を選択する選択手段と、
     選択された前記推奨条件を出力する出力手段とを備える処理装置。
    The processing device according to claim 15,
    From the recommended combination information included in the recommended information, the sensor included in the information indicating the available sensor, the extraction unit that extracts the recommended combination information indicating the feasible combination,
    From the recommended conditions associated with the extracted recommended combination information, selecting means for selecting one or more of the recommended conditions,
    Output means for outputting the selected recommended condition.
  17.  請求項16に記載の処理装置において、
     前記推奨情報において、前記推奨条件および前記推奨組み合わせ情報には、当該推奨条件と当該推奨組み合わせ情報が示す前記組み合わせとを用いた場合の、におい成分に関する予測精度がさらに関連づけられており、
     前記選択手段は、前記予測精度に基づいて、一以上の前記推奨条件を選択する処理装置。
    The processing device according to claim 16,
    In the recommended information, the recommended conditions and the recommended combination information, when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated,
    The processing device, wherein the selecting unit selects one or more of the recommended conditions based on the prediction accuracy.
  18.  請求項16または17に記載の処理装置において、
     前記選択手段は、前記推奨条件の広さおよび予め定められた条件からの近さの少なくとも一方に基づいて、一以上の前記推奨条件を選択する処理装置。
    The processing device according to claim 16 or 17,
    The processing device, wherein the selecting unit selects one or more of the recommended conditions based on at least one of a size of the recommended condition and a proximity to a predetermined condition.
  19.  請求項11から18のいずれか一項に記載の処理装置において、
     前記検出環境は、温度、湿度、気圧、夾雑ガスの種類、パージガスの種類、サンプリング周期、対象物と前記センサとの距離、前記センサの周囲に存在する物体のうち少なくともいずれかを含む処理装置。
    The processing device according to any one of claims 11 to 18,
    The processing apparatus according to claim 1, wherein the detection environment includes at least one of a temperature, a humidity, an atmospheric pressure, a type of an impurity gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor.
  20.  複数種類のセンサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づき、前記センサの検出環境の推奨条件と、一以上の前記センサからなる組み合わせとを関連づけた推奨情報を生成する推奨情報生成ステップを含む情報処理方法。 Based on execution results of machine learning using a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct answer data, based on a recommended condition of a detection environment of the sensor and a combination of one or more of the sensors An information processing method that includes a recommended information generating step of generating recommended information in which recommended information is associated with.
  21.  請求項20に記載の情報処理方法において、
     前記機械学習を行うことにより、前記複数の特徴量を変数とする式であって、におい成分に関する予測を行うための予測式を生成する予測式生成ステップと、
     前記予測式における前記複数の特徴量に対する複数の重みに基づいて、前記集合から一以上の前記センサを抽出し、抽出した前記センサからなる前記組み合わせを示す推奨組み合わせ情報を生成する抽出ステップとをさらに含み、
     前記抽出ステップでは、前記予測式において、前記複数の重みのうち予め定められた条件を満たす、または満たさない前記重みで重みづけられた前記特徴量の、出力元である前記センサを抽出し、
     前記推奨情報は、一以上の前記推奨組み合わせ情報を含む情報処理方法。
    The information processing method according to claim 20,
    By performing the machine learning, a prediction formula generating step of generating a prediction formula for performing a prediction on an odor component, wherein the prediction formula generation step is a formula using the plurality of feature amounts as variables,
    Extracting one or more of the sensors from the set based on a plurality of weights for the plurality of feature amounts in the prediction formula, and generating recommended combination information indicating the combination of the extracted sensors. Including
    In the extracting step, in the prediction equation, satisfying a predetermined condition among the plurality of weights, or extracting the sensor as an output source of the feature amount weighted by the weight that does not satisfy,
    The information processing method, wherein the recommended information includes one or more pieces of the recommended combination information.
  22.  請求項21に記載の情報処理方法において、
     前記予測式生成ステップでは、
      前記検出環境に基づいた分岐を含むモデルを用いて前記予測式を生成し、
      前記推奨組み合わせ情報に、前記予測式に適した前記検出環境の条件であって、前記分岐の条件に基づく前記検出環境の条件を前記推奨条件として関連づける情報処理方法。
    The information processing method according to claim 21,
    In the prediction formula generation step,
    Using the model including a branch based on the detection environment to generate the prediction formula,
    An information processing method for associating the recommended combination information with a condition of the detection environment suitable for the prediction equation, the condition of the detection environment based on the condition of the branch, as the recommended condition.
  23.  請求項22に記載の情報処理方法において、
     前記機械学習は、前記特徴量に関連づけられた前記検出環境をさらに入力とした異種混合学習であり、
     前記分岐の条件は、前記異種混合学習により生成される情報処理方法。
    The information processing method according to claim 22,
    The machine learning is a heterogeneous mixture learning further inputting the detection environment associated with the feature amount,
    The information processing method according to claim 1, wherein the branch condition is generated by the heterogeneous mixture learning.
  24.  請求項21から23のいずれか一項に記載の情報処理方法において、
     前記検出環境は、温度、湿度、気圧、夾雑ガスの種類、パージガスの種類、サンプリング周期、対象物と前記センサとの距離、前記センサの周囲に存在する物体のうち少なくともいずれかを含む情報処理方法。
    The information processing method according to any one of claims 21 to 23,
    The information processing method includes at least one of a temperature, a humidity, an atmospheric pressure, a type of an impurity gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor. .
  25.  請求項21から24のいずれか一項に記載の情報処理方法において、
     前記予測式の予測精度を算出する予測精度算出ステップをさらに含み、
     前記推奨情報は、前記推奨組み合わせ情報に関連づけられた前記予測式の前記予測精度をさらに含む情報処理方法。
    The information processing method according to any one of claims 21 to 24,
    The method further includes a prediction accuracy calculation step of calculating the prediction accuracy of the prediction expression,
    The information processing method, wherein the recommendation information further includes the prediction accuracy of the prediction formula associated with the recommendation combination information.
  26.  請求項21から25のいずれか一項に記載の情報処理方法において、
     前記組み合わせを採用する場合のコストに少なくとも基づいて、前記組み合わせを評価する評価ステップをさらに含み、
     前記推奨情報は、前記推奨組み合わせ情報に関連づけられた、前記評価ステップでの評価結果をさらに含む情報処理方法。
    The information processing method according to any one of claims 21 to 25,
    At least based on the cost of employing the combination, further comprising an evaluation step of evaluating the combination,
    The information processing method, wherein the recommendation information further includes an evaluation result in the evaluation step, which is associated with the recommendation combination information.
  27.  請求項26に記載の情報処理方法において、
     前記評価ステップでは、前記推奨組み合わせ情報に関連づけられた前記推奨条件にさらに基づいて、前記組み合わせを評価する情報処理方法。
    The information processing method according to claim 26,
    In the evaluating step, an information processing method for evaluating the combination based on the recommended condition associated with the recommended combination information.
  28.  センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、前記検出環境を示す情報とに基づいて、前記組み合わせを出力し、
     前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である処理方法。
    Based on information indicating the detection environment and the recommended information relating the recommended conditions of the sensor detection environment and the recommended combination information indicating the combination of one or more sensors, and outputting the combination,
    A processing method, wherein the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data.
  29.  請求項28に記載の処理方法において、
     前記推奨情報に含まれる前記推奨条件から、前記検出環境を示す情報が適合する前記推奨条件を抽出する抽出ステップと、
     抽出された前記推奨条件に関連づけられた前記推奨組み合わせ情報から、一以上の前記推奨組み合わせ情報を選択する選択ステップと、
     選択された前記推奨組み合わせ情報が示す前記組み合わせを出力する出力ステップとを含む処理方法。
    The processing method according to claim 28,
    From the recommended conditions included in the recommended information, an extraction step of extracting the recommended conditions to which information indicating the detection environment is suitable,
    A selection step of selecting one or more of the recommended combination information from the recommended combination information associated with the extracted recommended condition,
    Outputting the combination indicated by the selected recommended combination information.
  30.  請求項29に記載の処理方法において、
     前記推奨情報において、前記推奨条件および前記推奨組み合わせ情報には、当該推奨条件と当該推奨組み合わせ情報が示す前記組み合わせとを用いた場合の、におい成分に関する予測精度がさらに関連づけられており、
     前記選択ステップでは、抽出された前記推奨条件に関連づけられた前記推奨組み合わせ情報から、前記予測精度に基づいて、一以上の前記推奨組み合わせ情報を選択する処理方法。
    The processing method according to claim 29,
    In the recommended information, the recommended conditions and the recommended combination information, when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated,
    In the selecting step, a processing method of selecting one or more pieces of the recommended combination information from the recommended combination information associated with the extracted recommended condition based on the prediction accuracy.
  31.  請求項29または30に記載の処理方法において、
     前記選択ステップでは、前記推奨組み合わせ情報が示す前記組み合わせを用いた場合の、コストに基づいて、一以上の前記推奨組み合わせ情報を選択する処理方法。
    The processing method according to claim 29 or 30,
    In the selecting step, a processing method for selecting one or more pieces of the recommended combination information based on a cost when the combination indicated by the recommended combination information is used.
  32.  センサの検出環境の推奨条件と一以上の前記センサからなる組み合わせを示す推奨組み合わせ情報とを関連づけた推奨情報と、使用可能な前記センサを示す情報とに基づいて、前記推奨条件を出力し、
     前記推奨情報は、複数種類の前記センサの集合からの出力に基づいた複数の特徴量と正解データとを入力とした機械学習の実行結果に基づく情報である処理方法。
    Based on information indicating the recommended conditions of the detection environment of the sensor and recommended combination information indicating a combination of one or more sensors, and information indicating the available sensors, the recommended conditions are output.
    A processing method, wherein the recommendation information is information based on an execution result of machine learning that receives a plurality of feature amounts based on outputs from a set of a plurality of types of sensors and correct data.
  33.  請求項32に記載の処理方法において、
     前記推奨情報に含まれる前記推奨組み合わせ情報から、前記使用可能な前記センサを示す情報に含まれる前記センサで、実現可能な前記組み合わせを示す前記推奨組み合わせ情報を抽出する抽出ステップと、
     抽出された前記推奨組み合わせ情報に関連づけられた前記推奨条件から、一以上の前記推奨条件を選択する選択ステップと、
     選択された前記推奨条件を出力する出力ステップとを含む処理方法。
    The processing method according to claim 32,
    From the recommended combination information included in the recommended information, the sensor included in the information indicating the available sensor, an extraction step of extracting the recommended combination information indicating the feasible combination,
    A selecting step of selecting one or more of the recommended conditions from the recommended conditions associated with the extracted recommended combination information,
    Outputting the selected recommended condition.
  34.  請求項33に記載の処理方法において、
     前記推奨情報において、前記推奨条件および前記推奨組み合わせ情報には、当該推奨条件と当該推奨組み合わせ情報が示す前記組み合わせとを用いた場合の、におい成分に関する予測精度がさらに関連づけられており、
     前記選択ステップでは、前記予測精度に基づいて、一以上の前記推奨条件を選択する処理方法。
    The processing method according to claim 33,
    In the recommended information, the recommended conditions and the recommended combination information, when using the recommended conditions and the combination indicated by the recommended combination information, the prediction accuracy of the odor component is further associated,
    In the selecting step, a processing method of selecting one or more recommended conditions based on the prediction accuracy.
  35.  請求項33または34に記載の処理方法において、
     前記選択ステップでは、前記推奨条件の広さおよび予め定められた条件からの近さの少なくとも一方に基づいて、一以上の前記推奨条件を選択する処理方法。
    The processing method according to claim 33 or 34,
    In the selecting step, a processing method of selecting one or more of the recommended conditions based on at least one of a size of the recommended condition and a proximity from a predetermined condition.
  36.  請求項28から35のいずれか一項に記載の処理方法において、
     前記検出環境は、温度、湿度、気圧、夾雑ガスの種類、パージガスの種類、サンプリング周期、対象物と前記センサとの距離、前記センサの周囲に存在する物体のうち少なくともいずれかを含む処理方法。
    The processing method according to any one of claims 28 to 35,
    The processing method, wherein the detection environment includes at least one of a temperature, a humidity, an atmospheric pressure, a type of a contaminant gas, a type of a purge gas, a sampling cycle, a distance between an object and the sensor, and an object existing around the sensor.
  37.  請求項20から27のいずれか一項に記載の情報処理方法の各ステップをコンピュータに実行させるプログラム。 A program for causing a computer to execute each step of the information processing method according to any one of claims 20 to 27.
  38.  請求項28から36のいずれか一項に記載の処理方法の各ステップをコンピュータに実行させるプログラム。 A program for causing a computer to execute each step of the processing method according to any one of claims 28 to 36.
PCT/JP2018/035851 2018-09-27 2018-09-27 Information processing device, processing device, information processing method, processing method, determination method, and program WO2020065807A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/JP2018/035851 WO2020065807A1 (en) 2018-09-27 2018-09-27 Information processing device, processing device, information processing method, processing method, determination method, and program
JP2020547702A JP7056747B2 (en) 2018-09-27 2018-09-27 Information processing equipment, processing equipment, information processing method, processing method, determination method, and program

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2018/035851 WO2020065807A1 (en) 2018-09-27 2018-09-27 Information processing device, processing device, information processing method, processing method, determination method, and program

Publications (1)

Publication Number Publication Date
WO2020065807A1 true WO2020065807A1 (en) 2020-04-02

Family

ID=69949779

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2018/035851 WO2020065807A1 (en) 2018-09-27 2018-09-27 Information processing device, processing device, information processing method, processing method, determination method, and program

Country Status (2)

Country Link
JP (1) JP7056747B2 (en)
WO (1) WO2020065807A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2023007937A1 (en) * 2021-07-30 2023-02-02 三菱電機株式会社 Air conditioning control device and air conditioning control method

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2018109876A (en) * 2017-01-04 2018-07-12 株式会社東芝 Sensor design support apparatus, sensor design support method and computer program
WO2018150798A1 (en) * 2017-02-17 2018-08-23 日本電気株式会社 Model estimation system, method, and program
JP2018132325A (en) * 2017-02-13 2018-08-23 国立研究開発法人物質・材料研究機構 Non-destructive inspection method and device for degree of ripening of western pear by smell measurement

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018101128A1 (en) * 2016-11-29 2018-06-07 国立研究開発法人物質・材料研究機構 Method and device for estimating value to be estimated associated with specimen

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2018109876A (en) * 2017-01-04 2018-07-12 株式会社東芝 Sensor design support apparatus, sensor design support method and computer program
JP2018132325A (en) * 2017-02-13 2018-08-23 国立研究開発法人物質・材料研究機構 Non-destructive inspection method and device for degree of ripening of western pear by smell measurement
WO2018150798A1 (en) * 2017-02-17 2018-08-23 日本電気株式会社 Model estimation system, method, and program

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2023007937A1 (en) * 2021-07-30 2023-02-02 三菱電機株式会社 Air conditioning control device and air conditioning control method
JP7455092B2 (en) 2021-07-30 2024-03-25 三菱電機株式会社 Air conditioning control device and air conditioning control method

Also Published As

Publication number Publication date
JPWO2020065807A1 (en) 2021-09-09
JP7056747B2 (en) 2022-04-19

Similar Documents

Publication Publication Date Title
WO2020065806A1 (en) Processing device, processing method, and program
Wang et al. Optimal subsampling for quantile regression in big data
CN106251174A (en) Information recommendation method and device
Sansó et al. Statistical calibration of climate system properties
JP7267883B2 (en) Material property prediction system and material property prediction method
WO2020065807A1 (en) Information processing device, processing device, information processing method, processing method, determination method, and program
WO2020255305A1 (en) Prediction model re-learning device, prediction model re-learning method, and program recording medium
Andreu et al. Prediction of bioactive peptides using artificial neural networks
Asaadi et al. The use of direct inverse maps to solve material identification problems: pitfalls and solutions
US20210134397A1 (en) Information processing apparatus, information processing method, and storage medium
WO2020065991A1 (en) Information processing device, sensor operation optimization method, and program
JP7140191B2 (en) Information processing device, control method, and program
JP7127697B2 (en) Information processing device, control method, and program
WO2020026328A1 (en) Information processing device, control method, and program
CN110851587B (en) Commodity coding prediction model generation and commodity coding determination method, device and equipment
Wang et al. Nonparametric interval estimators for the coefficient of variation
JP7099623B2 (en) Information processing equipment, information processing methods, and programs
JP7074194B2 (en) Information processing equipment, control methods, and programs
JP7099296B2 (en) Evaluation method, system construction method, and evaluation system
CN118043802A (en) Recommendation model training method and device
JP2022167027A (en) Model generation device, model generation method, model generation program, quality prediction model and product quality prediction device
JP7327648B2 (en) DATA GENERATION DEVICE, DATA GENERATION METHOD, AND PROGRAM
WO2020065983A1 (en) Information processing device, control method, and program
Park et al. Assessing the Multi-labelness of Multi-label Data
JP2022127316A (en) Physical property predictor and program

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18935317

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2020547702

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18935317

Country of ref document: EP

Kind code of ref document: A1