WO2023038363A1 - Rhinitis diagnosis apparatus, method, and recording medium - Google Patents

Rhinitis diagnosis apparatus, method, and recording medium Download PDF

Info

Publication number
WO2023038363A1
WO2023038363A1 PCT/KR2022/013025 KR2022013025W WO2023038363A1 WO 2023038363 A1 WO2023038363 A1 WO 2023038363A1 KR 2022013025 W KR2022013025 W KR 2022013025W WO 2023038363 A1 WO2023038363 A1 WO 2023038363A1
Authority
WO
WIPO (PCT)
Prior art keywords
rhinitis
variable
model
information
patient
Prior art date
Application number
PCT/KR2022/013025
Other languages
French (fr)
Korean (ko)
Inventor
이화영
Original Assignee
가톨릭대학교 산학협력단
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 가톨릭대학교 산학협력단 filed Critical 가톨릭대학교 산학협력단
Publication of WO2023038363A1 publication Critical patent/WO2023038363A1/en

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/08—Detecting, measuring or recording devices for evaluating the respiratory organs
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01W—METEOROLOGY
    • G01W1/00—Meteorology
    • G01W1/02—Instruments for indicating weather conditions by measuring two or more variables, e.g. humidity, pressure, temperature, cloud cover or wind speed
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/20—Ensemble learning
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients

Definitions

  • the present embodiments provide rhinitis diagnosis apparatus, method and recording medium.
  • rhinitis can be diagnosed relatively easily because it has characteristic symptoms according to the patient's medical history. Ask well about symptoms related to the nose, and also check the relationship to the environment or job, and the weather. Even in the same allergic rhinitis, since the causative substances are all different, it is important to diagnose the patient's medical history and determine the treatment according to the severity of symptoms and environmental exposure. but. This diagnosis has a problem in that patients can receive it only by visiting a hospital every day.
  • the present embodiments can provide rhinitis diagnosis terminals, methods and recording media, methods and recording media that predict reliable rhinitis scores for rhinitis diagnosis and treatment of patients.
  • rhinitis diagnosis device significant variables are extracted from the patient's characteristic information including rhinitis severity information, environmental information and weather information through correlation analysis
  • a variable determining unit that determines a predictor variable, a model determining unit that creates a predictive model for rhinitis diagnosis based on the predictor variable, determines the predictive performance of the predictive model, and determines an optimized predictive model, and a patient corresponding to the predictor variable. It provides a rhinitis diagnostic device including a score prediction unit for predicting the patient's rhinitis score by inputting feature information into an optimized predictive model.
  • this embodiment is a method for managing rhinitis, a variable determination step of extracting a significant variable through correlation analysis from the patient's characteristic information including rhinitis severity information, environmental information and weather information and determining it as a predictor variable; A model decision step of generating a predictive model for rhinitis diagnosis based on the predictor variables and determining the optimized predictive model by determining the predictive performance of the predictive model and inputting the patient's characteristic information corresponding to the predictor variable into the optimized predictive model To provide a rhinitis diagnosis method comprising a score prediction step of predicting the patient's rhinitis score.
  • this embodiment extracts significant variables through correlation analysis from patient characteristic information including rhinitis severity information, environmental information, and weather information in a recording medium recording a program for executing a method for diagnosing rhinitis.
  • Patients corresponding to predictor variables and model determination function that generates a predictive model for rhinitis diagnosis based on the predictor variable, determines the predictive performance of the predictive model, and determines the optimized predictive model.
  • FIG. 1 is a diagram exemplarily illustrating a system configuration to which the present disclosure may be applied.
  • FIG. 2 is a diagram showing the configuration of a rhinitis diagnosis device according to an embodiment of the present disclosure.
  • Figure 3 is a flow chart for explaining the operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure to predict rhinitis score.
  • Figure 4 is a diagram for explaining the operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure determines the predictive variable.
  • Figure 5 is a diagram for explaining the operation of rhinitis diagnosis apparatus according to another embodiment of the present disclosure determines the predictive variable.
  • FIG. 6 is a diagram for explaining an operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure using a slope of weather information.
  • FIG. 7 is a diagram showing an example of inputting characteristic information of a patient in the rhinitis diagnosis apparatus according to an embodiment of the present disclosure.
  • FIG. 8 is a diagram showing an example of using weather information in the rhinitis diagnosis device according to an embodiment of the present disclosure.
  • FIG. 9 is a diagram showing an example of predicting a rhinitis score in the rhinitis diagnostic device according to an embodiment of the present disclosure.
  • FIG. 10 is a flow chart of a method for diagnosing rhinitis according to an embodiment of the present disclosure.
  • FIG. 11 is a diagram conceptually illustrating the configuration of a recording medium according to an embodiment of the present disclosure.
  • the present disclosure relates to a device for diagnosing rhinitis, a method, and a recording medium.
  • first, second, A, B, (a), and (b) may be used in describing the components of the present disclosure. These terms are only used to distinguish the component from other components, and the nature, order, or order of the corresponding component is not limited by the term.
  • an element is described as being “connected,” “coupled to,” or “connected” to another element, that element is directly connected or connectable to the other element, but there is another element between the elements. It will be understood that elements may be “connected”, “coupled” or “connected”.
  • VAS Visual Analogue Scale
  • symptom information in the present specification may mean a visual analog score in which the patient self-diagnoses the severity of nasal symptoms and checks the degree
  • SNOT-22 Tino-Nasal Outcome Test
  • Rhinitis score in the present specification is a result predicted through a predictive model based on a patient's symptom score, and may mean a score consistent with rhinitis diagnosis-related variables in a hospital.
  • the rhinitis score in the present specification may be used in the same sense as the VAS nose score (Easysum).
  • FIG. 1 is a diagram exemplarily illustrating a system configuration to which the present disclosure may be applied.
  • the present disclosure relates to a system for providing a rhinitis diagnosis device, may be implemented in the rhinitis diagnosis device 100 and the server (110).
  • Rhinitis diagnosis device 100 may include a general PC such as a general desktop or laptop, and may include a mobile terminal such as a smart phone, a tablet PC, a PDA (Personal Digital Assistants) and a mobile communication terminal, and is limited thereto. It is not, and should be broadly interpreted as any electronic device capable of communicating with the server 110.
  • a general PC such as a general desktop or laptop
  • a mobile terminal such as a smart phone, a tablet PC, a PDA (Personal Digital Assistants) and a mobile communication terminal, and is limited thereto. It is not, and should be broadly interpreted as any electronic device capable of communicating with the server 110.
  • the server 110 has the same configuration as a conventional web server (Web Server), web application server (Web Application Server), or web server (WAP Server) in terms of hardware.
  • Web Server web server
  • Web Application Server web application server
  • WAP Server web server
  • program modules that are implemented through any language such as C, C++, Java, PHP, .Net, Python, Ruby, and perform various functions. can do.
  • the server 110 may be connected to an unspecified number of clients (including the rhinitis diagnosis device 100) and/or other servers through a network. Accordingly, the server 110 receives requests from clients or other servers to perform tasks. It may mean a computer system that accepts and derives and provides work results for it, or computer software (server program) installed for such a computer system.
  • server 110 is understood as a broad concept including, in addition to the above-described server program, a series of application programs that operate on the server 110 and, in some cases, various databases built inside or outside. It should be.
  • the database may refer to an aggregate of data in which data such as information or data is structured and managed for use by a server or other device, and may also refer to a storage medium for storing such an aggregate of data.
  • a database may include a plurality of databases classified according to a data structure method, management method, type, and the like.
  • the database may include a database management system (DBMS), which is software that allows information or data to be added, corrected, or deleted.
  • DBMS database management system
  • the server 110 may store and manage contents and various types of information and data in a database.
  • the database may be implemented inside or outside the server 110 .
  • the server 110 uses server programs that are provided in various ways according to operating systems such as DOS, Windows, Linux, UNIX, and Macintosh in general server hardware. It can be implemented, and as a representative example, a website, IIS (Internet Information Server) used in a Windows environment, and Apache, Nginx, Light HTTP, etc. used in a Unix environment can be used.
  • operating systems such as DOS, Windows, Linux, UNIX, and Macintosh in general server hardware. It can be implemented, and as a representative example, a website, IIS (Internet Information Server) used in a Windows environment, and Apache, Nginx, Light HTTP, etc. used in a Unix environment can be used.
  • IIS Internet Information Server
  • the network 120 is a network that connects the server 110 and the rhinitis diagnosis device 100, and may be a closed network such as a LAN (Local Area Network) and a WAN (Wide Area Network), but the Internet It may be an open network such as (Internet).
  • the Internet refers to the TCP/IP protocol and various services existing in its upper layer, namely HTTP (HyperText Transfer Protocol), Telnet, FTP (File Transfer Protocol), DNS (Domain Name System), SMTP (Simple Mail Transfer Protocol), It refers to a worldwide open computer network structure that provides Simple Network Management Protocol (SNMP), Network File Service (NFS), and Network Information Service (NIS).
  • SNMP Simple Network Management Protocol
  • NFS Network File Service
  • NIS Network Information Service
  • the rhinitis diagnosis device 100 includes a mobile terminal such as a smart phone, a tablet PC, a PDA (Personal Digital Assistants) and a mobile communication terminal
  • the network is a wireless access network such as a mobile communication network or a Wi-Fi (WiFi) network. may include more.
  • Rhinitis diagnosis apparatus 100 in the present specification may refer to a terminal of a user (User) that predicts a rhinitis score by receiving characteristic information of a patient in order to diagnose rhinitis of a patient.
  • the rhinitis diagnosis device 100 may refer to a device used by a user who has been granted permission to access the predictive model determined by the server 110.
  • the rhinitis diagnosis apparatus 100 may refer to a user device that obtains the patient's characteristic information described later and transmits it to the server.
  • the rhinitis diagnosis device 100 in the present specification may connect to the server 110 to upload / download information for diagnosing rhinitis to the content platform.
  • the content platform may mean an online platform capable of predicting rhinitis scores operated or operated by the server 110 .
  • FIG. 2 is a diagram showing the configuration of a rhinitis diagnosis device according to an embodiment of the present disclosure.
  • the rhinitis diagnosis apparatus 100 which provides a method for diagnosing rhinitis according to an embodiment of the present disclosure, is significant through correlation analysis from patient characteristic information including rhinitis severity information, environmental information, and weather information.
  • a variable determining unit 210 that extracts a variable and determines it as a predictor variable, a model determining unit that generates a predictive model for diagnosing rhinitis based on the predictor variable, and determines the predictive performance of the predictive model to determine an optimized predictive model.
  • the variable determination unit 210 may extract significant variables through correlation analysis from the patient's characteristic information including rhinitis severity information, environmental information, and weather information, and determine them as predictive variables.
  • the rhinitis severity information may include at least one of SNOT-22 (Sino-Nasal Outcome Test) information, information on the presence or absence of disturbances in daily life, and VAS symptom information.
  • environmental information may include information on mold, transportation, stress, or cold air.
  • the variable determining unit 210 may receive input of patient characteristic information including rhinitis severity information and environmental information from the patient through an input interface.
  • the input interface may refer to a module or device capable of inputting information to the rhinitis diagnosis device 100, such as a touch screen, a microphone, and a button.
  • the variable determiner 210 may obtain VAS symptom information, which is self-diagnosis information about nasal symptoms input from a patient, as rhinitis severity information.
  • the variable determiner 210 may obtain information on a symptom aggravating factor input from the patient as environmental information.
  • variable determining unit 210 may receive SNOT-22 information of a patient diagnosed in a hospital from a server and obtain it as rhinitis severity information.
  • the variable determiner 210 may receive and obtain weather information according to the location information of the patient from the server.
  • the variable determiner 210 may receive temperature information, humidity information, and fine dust information of a current location using a global positioning system (GPS) to obtain weather information.
  • GPS global positioning system
  • variable determining unit 210 may determine a predictor variable suitable for the predictive model by dividing the significant variable extracted through correlation analysis with the rhinitis score of the patient into a common variable or an individual variable.
  • the variable determiner 210 may calculate a correlation coefficient for each variable of the patient's characteristic information, and classify significant variables extracted based on the correlation coefficient as common variables.
  • the common variable may mean a variable having a fixed effect.
  • the variable determination unit 210 may calculate a correlation coefficient (p) through correlation analysis (Pearson Correlation analysis) with the rhinitis score of the patient.
  • variable determining unit 210 may extract significant variables from variables having a distinct quantitative linear relationship such that the calculated correlation coefficient (p) is 0.3 or more and 0.7 or less, and classify them as common variables.
  • the variable determiner 210 may classify each of the temperature information, humidity information, and fine dust information included in the obtained weather information based on the location information of the patient as individual variables.
  • the individual variable may mean a variable having an arbitrary effect.
  • the slope of each of the temperature information, humidity information, and fine dust information for the rhinitis score may be different for each patient.
  • the variable determiner 210 may classify a random intercept as an individual variable.
  • variable determiner 210 may classify arbitrary segments as individual variables in order to consider the basic characteristics of each patient by reflecting the correlation as the distribution of rhinitis scores shows a difference between patients. This is because the predicted rhinitis score has different subjective characteristics for each individual.
  • the model determination unit 220 may generate a predictive model for diagnosing rhinitis based on predictor variables, and determine an optimized predictive model by determining the predictive performance of each predictive model.
  • the model determiner 220 may generate a plurality of predictive models including a regression model, a linear mixed model, and an ensemble machine learning model based on predictor variables. there is.
  • the model determination unit 220 may determine the prediction performance of each prediction model and determine an optimized prediction model from among a plurality of prediction models.
  • the model determiner 220 may generate a regression model that does not reflect any effect (individual variable), a linear mixed model that reflects only an arbitrary intercept, or a linear mixed model that reflects both the random intercept and the slope as a prediction model. .
  • the model determiner 220 may generate an ensemble model based on a random forest algorithm or an ensemble model based on a LightGBM algorithm as a predictive model.
  • each ensemble model may be composed of a model that does not reflect any effect (individual variable), a model that reflects only an arbitrary intercept, or a model that reflects both the random intercept and the slope.
  • the generated model is an example, and is not limited thereto.
  • the model determiner 220 may determine the predictive performance of the predictive model based on the root mean square error calculated by applying K-fold cross-validation to the patient's feature information corresponding to the predictor variable.
  • the model determination unit 220 may determine the degree of fitness of the model using Bayesian information criterion (BIC).
  • BIC Bayesian information criterion
  • the Bayes information criterion may be a numerical value used in Bayesian statistics as a criterion for selecting a model from among a plurality of models.
  • the model determiner 220 may determine an optimized prediction model using a root mean square error calculated by applying K-fold cross-validation to each of a plurality of prediction models and a Bayes information criterion.
  • the model determiner 220 may determine an ensemble model generated by applying a common variable and an individual variable to a tree-based machine learning algorithm as an optimized prediction model.
  • the score prediction unit 230 may predict the patient's rhinitis score by inputting the patient's characteristic information corresponding to the predictor variable into an optimized prediction model. For example, the score prediction unit 230 may obtain predictive validity by confirming the internal consistency of rhinitis diagnosis-related variables and the predicted rhinitis score through reliability analysis (Cronbach's alpha). At this time, variables related to rhinitis diagnosis may be variables related to SNOT-22 information performed for rhinitis diagnosis in a hospital.
  • Figure 3 is a flow chart for explaining the operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure to predict rhinitis score.
  • the variable determining unit 210 of the rhinitis diagnostic device may determine a predictor variable suitable for the predictive model by extracting a significant variable through a correlation analysis (S310). For example, the variable determination unit 210 may determine a predictor variable suitable for the predictive model by classifying the significant variable extracted through correlation analysis with the rhinitis score of the patient into a common variable or an individual variable. For example, the variable determination unit 210 may calculate a correlation coefficient (p) in the patient's rhinitis score and correlation analysis. Also, the variable determining unit 210 may extract a significant variable from variables having a distinct quantitative linear relationship such that the calculated correlation coefficient (p) is greater than or equal to 0.3 and less than or equal to 0.7.
  • a correlation coefficient (p) in the patient's rhinitis score and correlation analysis.
  • the variable determining unit 210 may extract a significant variable from variables having a distinct quantitative linear relationship such that the calculated correlation coefficient (p) is greater than or equal to 0.3 and less than or equal to 0.7.
  • the variable determination unit 210 may determine only significant variables as predictor variables by applying the linear mixed model among the extracted variables.
  • the correlation coefficient (p) calculated in the correlation analysis can be calculated as SNOT-22 information 0.66, VAS symptom score 0.45, information on the presence or absence of disturbance in daily life 0.36 included in rhinitis severity.
  • the correlation coefficient (p) can be calculated as mold 0.31, transportation 0.34, stress 0.36, and cold air 0.42 included in the environmental information.
  • the variable determining unit 210 may finally classify the SNOT-22 information, cold air, stress, and transportation information as common variables based on the correlation coefficient and determine them as predictive variables.
  • the correlation coefficient is also changed when patient data is changed.
  • variable determiner 210 may classify each of the temperature information, humidity information, and fine dust information included in the weather information as individual variables.
  • the variable determining unit 210 may have different slopes of temperature information, humidity information, and fine dust information for rhinitis scores for each patient. Accordingly, the weather information has a low correlation when considering the entire patient, but can be divided into individual variables having random effects as they are offset by different slopes. Also, the weather information may be applied to a prediction model by assuming a random slope.
  • variable determiner 210 may classify a random intercept as an individual variable.
  • the variable determiner 210 may classify the arbitrary intercept as an individual variable in order to consider the basic characteristics differently for each patient by reflecting the correlation. This is because each individual may have subjective characteristics as the patient's subjective rhinitis severity information is reflected in the predicted rhinitis score. In addition, since the distribution of rhinitis scores differs between patients, it is necessary to consider the characteristic information differently for each patient.
  • Model determination unit 220 of the rhinitis diagnosis device may generate a plurality of predictive models for rhinitis diagnosis based on the predictor variables (S320).
  • the model determiner 220 may generate a plurality of predictive models including a regression model, a linear mixed model, and an ensemble machine learning model based on predictor variables. there is.
  • the model determiner 220 may generate a regression model that does not reflect any effect (individual variable).
  • the VAS symptom information has a high correlation with the SNOT-22 information, a multicollinearity problem occurs and is insignificant, so it can be excluded from the predictor variables.
  • the model determination unit 220 is SNOT-22 information included in rhinitis severity information, information on the presence or absence of disturbances in daily life, information on mold, transportation, stress, cold air included in environmental information, and intercepts ( rintercept) as a predictor variable.
  • the model determiner 220 may generate a linear mixed model reflecting only the random intercept or a linear mixed model reflecting both the random intercept and the slope.
  • the model determining unit 220 may create a linear mixed model in consideration of the characteristic that the predictive rhinitis score reflects the subjectivity of the patient.
  • the linear mixed model is less significant than the regression model, but may be a reasonable model as an approach that assumes that each patient is different by reflecting the correlation within the patient.
  • the model determination unit 220 may generate a linear mixed model reflecting only an arbitrary intercept according to the rhinitis scores of different patient groups.
  • the variable correlated with the predicted rhinitis score is mainly environmental information, which can also be seen as a subjective factor that varies for each patient.
  • the information on fungi included in the environmental information is less significant in the corresponding linear mixed model, and considering that only a small number of patients entered as fungi as an aggravating factor, it can be excluded from the predictor variables. Therefore, the model determination unit 220 generates a regression model using the SNOT-22 information included in the rhinitis severity information, the information on transportation, stress, and cold air included in the environmental information, and the intercept as a predictor variable.
  • the model determination unit 220 may generate a linear mixed model reflecting both a random intercept and a slope as the slope of the weather information for the rhinitis score is different for each patient.
  • the model determiner 220 may generate an ensemble model based on a random forest algorithm or an ensemble model based on a LightGBM algorithm.
  • each ensemble model may be composed of a model that does not reflect any effect (individual variable), a model that reflects only an arbitrary intercept, or a model that reflects both the random intercept and the slope.
  • an ensemble model based on a random forest algorithm may be a model that outputs a final result by collecting classification or prediction results from a plurality of decision trees constructed in a training process.
  • the ensemble model based on the LightGBM algorithm may be a model using a 'leaf-wise' method of continuously splitting around a node having a maximum loss value in order to correct an error of a decision tree.
  • Model determining unit 220 of the rhinitis diagnosis apparatus may determine the predictive performance of each predictive model to determine the optimized predictive model (S330). For example, the model determiner 220 may determine the predictive performance of the predictive model based on the root mean square error calculated by applying K-fold cross-validation. For example, the model determiner 220 may calculate prediction performance k times by dividing the data into k pieces and dividing them into a training set and a validation set.
  • Overall RMSE may be a value obtained by taking the square root of K mean square errors calculated for each iteration. Table 1 shows the prediction performance of each prediction model judged using the overall RMSE.
  • the predictive performance of the predictive model differs depending on the type of random effect (individual variable) in the linear mixed model and the ensemble model.
  • the predictive performance of the model excluding random effects (individual variables) may be judged to be good.
  • the ensemble model it can be determined that the predictive performance of the model in which only the random intercept is reflected is good.
  • the model determiner 220 may determine the degree of fitness of the model using Bayesian information criterion (BIC). For example, the model determiner 220 may determine model fit using the adjusted coefficient of determination of the regression model.
  • the adjusted coefficient of determination is R 2 included in the output of the regression model, and may be an indicator for measuring whether the model explains past data well.
  • the resulting regression model calculated an adjusted coefficient of determination of 0.525, which could mean explaining 52.5% of the rhinitis score variance.
  • the linear mixed model reflecting only the random intercept can be calculated as 1826.423 based on the Bayes information criterion.
  • the linear mixed model that reflects both the random intercept and the slope is calculated as 1530.467, so it can be judged that the model fit is higher than the model that reflects only the random intercept.
  • Score prediction unit 230 of the rhinitis diagnosis device may predict the patient's rhinitis score using an optimized prediction model (S340).
  • the optimized prediction model may be an ensemble model generated by applying a tree-based machine learning algorithm to common variables and individual variables.
  • the predicted patient's rhinitis score may be a score for which consistency with variables determined from the patient's characteristic information is confirmed. The consistency reliability of the rhinitis score will be described later with reference to FIG. 9 .
  • Figure 4 is a diagram for explaining the operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure determines the predictive variable.
  • variable determination unit 210 of the rhinitis diagnosis device may explain the correlation analysis with the continuous variable.
  • the variable determining unit 210 may extract significant variables through correlation analysis of continuous variables included in the patient's rhinitis score and rhinitis severity information.
  • the variable determining unit 210 may calculate a correlation coefficient of 0.66 through correlation analysis between the rhinitis score and the SNOT-22 information.
  • the variable determining unit 210 may calculate a correlation coefficient of 0.45 through correlation analysis between the rhinitis score and the VAS symptom information.
  • the variable determiner 210 may calculate a correlation coefficient of 0.69 through correlation analysis between the SNOT-22 information and the VAS symptom information. Therefore, it can be confirmed that rhinitis score has a clear quantitative linear relationship with SNOT-22 information and VAS symptom information.
  • Figure 5 is a diagram for explaining the operation of rhinitis diagnosis apparatus according to another embodiment of the present disclosure determines the predictive variable.
  • variable determination unit 210 of the rhinitis diagnosis device may explain the correlation analysis with the discrete variable.
  • the variable determination unit 210 may perform rhinitis score and correlation analysis for each discrete variable and display it as a box plot.
  • the box plot may be a graph expressing numerical data to quickly check the range and median of the data set.
  • the variable determining unit 210 may calculate a correlation coefficient of 0.36 through correlation analysis with rhinitis scores and information about the presence or absence of disturbances in daily life input from the patient.
  • the variable determining unit 210 may calculate a correlation coefficient through a correlation analysis between the rhinitis score and each environmental information input from the patient.
  • variable determining unit 210 may calculate a correlation coefficient between the rhinitis score and mold as 0.31.
  • the variable determining unit 210 may calculate a correlation coefficient between the rhinitis score and transportation as 0.34.
  • Variable determining unit 210 may calculate the correlation coefficient between rhinitis score and stress as 0.36.
  • the variable determiner 210 may calculate a correlation coefficient between the rhinitis score and the cold air as 0.42. Therefore, it can be confirmed that the rhinitis score has a clear quantitative linear relationship with environmental information on symptom aggravating factors.
  • FIG. 6 is a diagram for explaining an operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure using a slope of weather information.
  • the variable determination unit 210 of the rhinitis diagnosis apparatus may explain the distribution and slope of rhinitis scores according to weather information for each patient.
  • the variable determining unit 210 may obtain weather information of a current location for each patient using a global navigation system (GPS) and calculate the distribution and slope of rhinitis scores according to the weather information.
  • the variable determining unit 210 may calculate the slope of temperature information, humidity information, and fine dust information for rhinitis scores for each patient. The calculated slope may have different values for each patient.
  • FIG. 6 shows the distribution and slope of rhinitis scores according to temperature information among weather information.
  • FIG. 7 is a diagram showing an example of inputting characteristic information of a patient in the rhinitis diagnosis apparatus according to an embodiment of the present disclosure.
  • the rhinitis severity information and environmental information from the rhinitis diagnosis apparatus 100 according to an embodiment of the present disclosure from the patient.
  • VAS symptom information 710 of the patient in the rhinitis diagnosis apparatus 100 is obtained by inputting the degree of symptom self-diagnosis for nasal symptoms by a patient, and may be information included in rhinitis severity information.
  • the VAS symptom information 710 may be acquired by inputting additional self-diagnosis information about asthma symptoms in addition to nasal symptoms.
  • the VAS symptom information 710 may be input by the patient at regular intervals, but collected and monitored for a specific period of time. The predetermined period may be a one-day interval, but is not limited thereto.
  • the information 720 on aggravating factors is obtained by selecting and acquiring a factor that is thought to be a cause of exacerbation of nasal symptoms by a patient, and may be information included in environmental information.
  • the information 720 on aggravating factors may include cold, fine dust, house dust, pets, mold, pollen, cold air, humidity, transportation, smell, smoking, stress, acid reflux, exercise, medication, and food. can However, it is not limited thereto.
  • FIG. 8 is a diagram showing an example of using weather information in the rhinitis diagnosis device according to an embodiment of the present disclosure.
  • weather information 810 of a patient may be obtained from the rhinitis diagnosis apparatus 100 according to an embodiment of the present disclosure.
  • the weather information 810 may be temperature information, humidity information, and fine dust (PM10) information for each patient obtained from the Korea Meteorological Administration based on a global navigation system (GPS) once a day.
  • GPS global navigation system
  • the weather information 810 can be obtained based on the location information detected using the global navigation system (GPS) of the rhinitis diagnosis apparatus 100 when a signal such as a screen touch is input from the patient.
  • FIG. 9 is a diagram showing an example of predicting a rhinitis score in the rhinitis diagnostic device according to an embodiment of the present disclosure.
  • the rhinitis score 910 may be a score output by inputting patient characteristic information into an optimized predictive model.
  • the rhinitis score 910 can confirm internal consistency with variables included in the patient's rhinitis severity information through reliability analysis (Cronbach's alpha). Accordingly, the predicted rhinitis score can obtain predictive validity as a value that can be replaced with rhinitis diagnosis.
  • the reliability coefficient is a value representing the reliability of the internal consistency of the test, and it can be analyzed whether items are composed of homogeneous elements.
  • the reliability coefficient may be calculated to be 0.7 or more. Accordingly, it can be confirmed that the consistency of four variables corresponding to rhinitis score 910, SNOT-22 information, information on the presence or absence of disturbances in daily life, and VAS symptom information is reliable.
  • the rhinitis diagnosis apparatus 100 may provide a self-management method according to the degree of control by utilizing the predicted patient's rhinitis score.
  • FIG. 10 is a flow chart of a method for diagnosing rhinitis according to an embodiment of the present disclosure.
  • the method for diagnosing rhinitis of the present disclosure may include a variable determination step of determining predictive variables (S1010).
  • the rhinitis diagnostic device may extract significant variables through correlation analysis from the patient's characteristic information including rhinitis severity information, environmental information, and weather information and determine them as predictive variables.
  • the rhinitis diagnosis apparatus may receive characteristic information of a patient including rhinitis severity information and environmental information from a patient through an input interface.
  • the rhinitis diagnosis apparatus may acquire VAS symptom information, which is self-diagnosis information about nasal symptoms input from a patient, as rhinitis severity information.
  • the variable determiner 210 may obtain information on a symptom aggravating factor input from the patient as environmental information.
  • the rhinitis diagnosis device may obtain rhinitis severity information by receiving SNOT-22 information of a patient diagnosed in a hospital from a server.
  • the rhinitis diagnosis apparatus may receive weather information according to the patient's location information from the server and obtain it as the patient's characteristic information.
  • the rhinitis diagnosis device may obtain weather information by receiving temperature information, humidity information, and fine dust information of the current location using a global navigation system (GPS).
  • GPS global navigation system
  • the rhinitis diagnosis apparatus may determine a predictor variable suitable for a predictive model by dividing a significant variable extracted through correlation analysis with a patient's rhinitis score into a common variable or an individual variable. For example, the rhinitis diagnosis device may calculate a correlation coefficient for each variable of the patient's characteristic information, and distinguish a significant variable extracted based on the calculated correlation coefficient as a common variable. Specifically, the rhinitis diagnosis device may calculate a correlation coefficient (p) from the patient's rhinitis score and correlation analysis (Pearson Correlation analysis).
  • the rhinitis diagnostic device can be divided into common variables by extracting significant variables from among variables having a clear quantitative linear relationship with a calculated correlation coefficient (p) of 0.3 or more and 0.7 or less.
  • the rhinitis diagnosis device may classify each of the temperature information, humidity information, and fine dust information included in the obtained weather information based on the patient's location information as individual variables.
  • Each of the temperature information, humidity information, and fine dust information included in the obtained weather information based on the patient's location information may be classified as individual variables.
  • a rhinitis diagnostic device may classify a random intercept as an individual variable.
  • Rhinitis diagnosis method of the present disclosure may include a model determination step of determining a predictive model (S1020).
  • the rhinitis diagnosis device may generate a predictive model for diagnosing rhinitis based on a predictor variable, and determine an optimized predictive model by determining the predictive performance of each predictive model.
  • the rhinitis diagnosis device may generate a plurality of predictive models including a regression model, a linear mixed model, and an ensemble model (Ensemble Machine Learning Model) based on predictor variables. Then, the rhinitis diagnostic device may determine the prediction performance of each predictive model to determine an optimized predictive model from among a plurality of predictive models.
  • the rhinitis diagnosis device may determine the predictive performance of the predictive model based on the root mean square error calculated by applying K-fold cross validation to the patient's characteristic information corresponding to the predictor variable.
  • rhinitis diagnosis device may determine the model fit using the Bayesian information criterion (BIC).
  • BIC Bayesian information criterion
  • the rhinitis diagnosis device may determine a predictive model optimized for a predictive model having a low value of the root mean square error and Bayesian information criterion calculated by applying K-fold cross-validation to each of a plurality of predictive models.
  • the rhinitis diagnosis device may determine an ensemble model generated by applying a common variable and an individual variable to a tree-based machine learning algorithm as an optimized predictive model.
  • Rhinitis diagnosis method of the present disclosure may include a score prediction step of predicting the rhinitis score (S1030).
  • the rhinitis diagnostic device may predict the patient's rhinitis score by inputting the patient's characteristic information corresponding to the predictor variable into an optimized predictive model.
  • the rhinitis diagnosis device can obtain predictive validity by confirming the internal consistency between the rhinitis diagnosis-related variables and the predicted rhinitis score through reliability analysis (Cronbach's alpha).
  • FIG. 11 is a diagram conceptually illustrating the configuration of a recording medium according to an embodiment of the present disclosure.
  • the recording medium 1100 recording the program for executing the method for diagnosing rhinitis extracts significant variables through correlation analysis from patient characteristic information including rhinitis severity information, environmental information, and weather information, and predicts variables.
  • the variable determining function 1110 may extract a significant variable through correlation analysis from the patient's characteristic information including rhinitis severity information, environmental information, and weather information and determine it as a predictor variable.
  • the variable determining function 1110 may determine a predictor variable suitable for a predictive model by classifying a significant variable extracted through correlation analysis with a patient's rhinitis score into a common variable or an individual variable.
  • the variable determination function 1110 may calculate a correlation coefficient for each variable of the patient's characteristic information, and classify significant variables extracted based on the correlation coefficient as common variables.
  • variable determination function 1110 classifies each of the temperature information, humidity information, and fine dust information included in the obtained weather information based on the patient's location information as individual variables, each temperature information for the rhinitis score, The slope of the humidity information and the fine dust information may be different for each patient.
  • the model determining function 1120 may generate a predictive model for diagnosing rhinitis based on predictor variables and determine an optimized predictive model by determining predictive performance of the predictive model.
  • the model determination function 1120 may generate a plurality of predictive models including a regression model, a linear mixed model, and an ensemble model based on predictor variables.
  • the model determination function 1120 may determine the prediction performance of each prediction model and determine an optimized prediction model from among a plurality of prediction models.
  • the model determination function 1120 may determine the predictive performance of the predictive model based on the root mean square error calculated by applying K-fold cross-validation to the patient's feature information corresponding to the predictor variable.
  • the model determination function 1120 may determine an ensemble model generated by applying a common variable and an individual variable to a tree-based machine learning algorithm as an optimized predictive model.
  • the rhinitis diagnosis method according to an embodiment of the present disclosure described above is implemented as an application (ie, a program) installed by default in the rhinitis diagnosis device 100 or directly installed by a user, and is readable by a computer such as the rhinitis diagnosis device 100. can be recorded on a recordable medium.
  • a program implementing the method for diagnosing rhinitis according to an embodiment of the present disclosure executes a variable determination function, a model determination function, a score prediction function, and the like. These programs can be recorded on a computer-readable recording medium and executed by a computer to execute the aforementioned functions.
  • the above-described program is a computer such as C, C ++, JAVA, machine language, etc. that the processor (CPU) of the computer can read. It may include code coded in a language.
  • These codes may include functional codes related to functions defining the above-described functions, and may include control codes related to execution procedures necessary for a processor of a computer to execute the above-described functions according to a predetermined procedure.
  • these codes may further include memory reference related codes for which location (address address) of the computer's internal or external memory should be referenced for additional information or media necessary for the computer's processor to execute the above-mentioned functions. .
  • the code is used by the computer processor to communicate with the computer's communication module (e.g., wired and/or wireless communication module). ) may further include communication-related codes for how to communicate with any other remote computer or server, and what information or media should be transmitted/received during communication.
  • the computer's communication module e.g., wired and/or wireless communication module.
  • a functional program for implementing the present disclosure codes and code segments related thereto, in consideration of the system environment of a computer that reads a recording medium and executes a program, etc. It may be easily inferred or changed by
  • the computer-readable recording medium on which the above-described program is recorded is distributed to computer systems connected through a network, so that computer-readable codes can be stored and executed in a distributed manner.
  • any one or more of the plurality of distributed computers may execute some of the functions presented above, transmit the execution results to one or more of the other distributed computers, and receive the transmitted results.
  • a computer may also execute some of the functions presented above and provide the results to other distributed computers as well.
  • a computer-readable recording medium recording a program for executing the rhinitis diagnosis method is, for example, ROM, RAM, CD-ROM, magnetic tape, floppy disk, and optical media storage devices.
  • the computer-readable recording medium recording the application which is a program for executing the rhinitis diagnosis method according to an embodiment of the present disclosure, is an application store server (Application Store Server), an application or a web server related to the service (Web Server ), etc., may be a storage medium (eg, hard disk, etc.) included in the application providing server (Application Provider Server), the application providing server itself, or another computer on which a program is recorded or its storage medium.
  • Application Store Server Application Store Server
  • Web Server web server
  • Storage medium eg, hard disk, etc.
  • a computer capable of reading a recording medium on which an application, which is a program for executing the rhinitis diagnosis method according to an embodiment of the present disclosure, is recorded, as well as a general PC such as a general desktop or laptop computer, a smart phone, a tablet PC, a PDA (Personal It may include mobile terminals such as Digital Assistants and mobile communication terminals, and should be interpreted as all devices capable of computing.
  • a general PC such as a general desktop or laptop computer
  • a smart phone such as a tablet PC
  • PDA Personal It may include mobile terminals such as Digital Assistants and mobile communication terminals, and should be interpreted as all devices capable of computing.
  • a mobile terminal such as a smart phone, tablet PC, PDA (Personal Digital Assistants) and mobile communication terminal
  • the mobile terminal may download and install the corresponding application from an application providing server including an application store server, a web server, etc.
  • the mobile device is downloaded through a synchronization program. It can also be installed in a terminal.

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Pathology (AREA)
  • Data Mining & Analysis (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • Veterinary Medicine (AREA)
  • Biophysics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • Theoretical Computer Science (AREA)
  • Environmental & Geological Engineering (AREA)
  • Software Systems (AREA)
  • Physiology (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Engineering & Computer Science (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Evolutionary Computation (AREA)
  • Atmospheric Sciences (AREA)
  • Pulmonology (AREA)
  • Ecology (AREA)
  • Environmental Sciences (AREA)
  • Medical Treatment And Welfare Office Work (AREA)
  • Immunology (AREA)
  • Vascular Medicine (AREA)
  • Psychiatry (AREA)
  • Signal Processing (AREA)

Abstract

The present disclosure relates to a rhinitis diagnosis apparatus, method, and recording medium, and can provide a rhinitis diagnosis apparatus, method, and recording medium, in which a rhinitis score is predicted by individually using characteristic information of a patient without the patient having to personally visit a hospital. In particular, provided are a rhinitis diagnosis apparatus, method, and recording medium for predicting a reliable rhinitis score by generating a prediction model for diagnosing rhinitis on the basis of a predictor variable extracted from characteristic information of a patient through a correlation analysis.

Description

비염 진단 장치, 방법 및 기록매체Rhinitis diagnosis device, method and recording medium
본 실시예들은 비염 진단 장치, 방법 및 기록매체를 제공한다.The present embodiments provide rhinitis diagnosis apparatus, method and recording medium.
산업 발전에 따른 대기오염에 의해 생활 환경이 변하고 있고, 야외생활의 증가나 새로운 환경 물질의 증가에 의해 알레르기성 비염이 매년 증가 추세로, 알레르기성 비염 환자는 이미 전 인구의 약 10 %정도로 추정된다. 이러한 증가 추세는 건강에 심각한 위협이 되고 있으며 환자 개인의 고통뿐만 아니라 경제적, 사회적 손실이 매우 커서 시급한 대책이 요구되고 있다. 알레르기성 비염은 기관지 천식만큼 심각한 건강상의 위협을 초래하지는 않으므로 관심이 적을 수 있으나, 유병율이 기관지 천식보다 매우 높으며 일상생활에서 심한 불편을 가져올 수 있어 최근에 주목을 받고 있다.The living environment is changing due to air pollution caused by industrial development, and allergic rhinitis is on the rise every year due to the increase in outdoor life or new environmental substances. . This increasing trend poses a serious threat to health, and an urgent countermeasure is required as the economic and social losses as well as the suffering of individual patients are very large. Allergic rhinitis may not cause a serious health threat as much as bronchial asthma, so there may be less interest, but it has a much higher prevalence than bronchial asthma and can cause severe inconvenience in daily life, so it has recently attracted attention.
또한, 비염의 진단은 환자의 병력에 따라 특징적인 증상을 가지고 있으므로 비교적 쉽게 진단할 수 있다. 코에 관련되는 증상을 잘 물어 보고 환경 또는 직업과의 관련성, 날씨에 대한 사항도 점검한다. 같은 알레르기성 비염일지라도 원인 물질은 제각기 모두 다르므로 환자의 병력을 문진하여 진단하고 증상의 중증도와 환경적 노출에 따라 치료제를 결정하는 것이 중요하다. 하지만. 이러한 진단은 환자들이 매일 병원을 방문해야만 받을 수 있다는 점에서 문제점이 있다.In addition, rhinitis can be diagnosed relatively easily because it has characteristic symptoms according to the patient's medical history. Ask well about symptoms related to the nose, and also check the relationship to the environment or job, and the weather. Even in the same allergic rhinitis, since the causative substances are all different, it is important to diagnose the patient's medical history and determine the treatment according to the severity of symptoms and environmental exposure. but. This diagnosis has a problem in that patients can receive it only by visiting a hospital every day.
따라서, 최근 디지털 의료의 발전에 따라 환자가 직접 병원을 방문하지 않더라도 모바일 기술을 사용하여 환자로부터 입력된 정보를 기반으로 진단하는 기술을 필요로 하고 있다. 특히, 환자가 간단하게 호흡기 알레르기 질환을 더 잘 예방하고 관리할 수 있도록 신뢰할 수 있는 진단 기술을 필요로 하고 있다.Therefore, with the recent development of digital medicine, there is a need for a technology for diagnosing based on information input from a patient using mobile technology even if the patient does not directly visit the hospital. In particular, there is a need for reliable diagnostic technology so that patients can simply better prevent and manage respiratory allergic diseases.
이러한 배경에서, 본 실시예들은 환자의 비염 진단과 치료를 위해 신뢰할 수 있는 비염 점수를 예측하는 비염 진단 단말, 방법 및 기록매체, 방법 및 기록매체를 제공할 수 있다.Against this background, the present embodiments can provide rhinitis diagnosis terminals, methods and recording media, methods and recording media that predict reliable rhinitis scores for rhinitis diagnosis and treatment of patients.
전술한 목적을 달성하기 위하여, 일 측면에서, 본 실시예는, 비염 진단 장치에 있어서, 비염 중증도 정보, 환경 정보 및 날씨 정보를 포함하는 환자의 특징 정보로부터 상관 분석을 통해 유의한 변수를 추출하여 예측 변수로 결정하는 변수 결정부, 예측 변수를 기반으로 비염 진단을 위한 예측 모델을 생성하고, 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정하는 모델 결정부 및 예측 변수에 해당되는 환자의 특징 정보를 최적화된 예측 모델에 입력하여 환자의 비염 점수를 예측하는 점수 예측부를 포함하는 비염 진단 장치를 제공한다.In order to achieve the above object, in one aspect, in the rhinitis diagnosis device, significant variables are extracted from the patient's characteristic information including rhinitis severity information, environmental information and weather information through correlation analysis A variable determining unit that determines a predictor variable, a model determining unit that creates a predictive model for rhinitis diagnosis based on the predictor variable, determines the predictive performance of the predictive model, and determines an optimized predictive model, and a patient corresponding to the predictor variable. It provides a rhinitis diagnostic device including a score prediction unit for predicting the patient's rhinitis score by inputting feature information into an optimized predictive model.
다른 측면에서, 본 실시예는 비염 관리 방법에 있어서, 비염 중증도 정보, 환경 정보 및 날씨 정보를 포함하는 환자의 특징 정보로부터 상관 분석을 통해 유의한 변수를 추출하여 예측 변수로 결정하는 변수 결정 단계, 예측 변수를 기반으로 비염 진단을 위한 예측 모델을 생성하고, 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정하는 모델 결정 단계 및 예측 변수에 해당되는 환자의 특징 정보를 최적화된 예측 모델에 입력하여 환자의 비염 점수를 예측하는 점수 예측 단계를 포함하는 비염 진단 방법을 제공한다.In another aspect, this embodiment is a method for managing rhinitis, a variable determination step of extracting a significant variable through correlation analysis from the patient's characteristic information including rhinitis severity information, environmental information and weather information and determining it as a predictor variable; A model decision step of generating a predictive model for rhinitis diagnosis based on the predictor variables and determining the optimized predictive model by determining the predictive performance of the predictive model and inputting the patient's characteristic information corresponding to the predictor variable into the optimized predictive model To provide a rhinitis diagnosis method comprising a score prediction step of predicting the patient's rhinitis score.
또 다른 측면에서, 본 실시예는 비염 진단 방법을 실행시키기 위한 프로그램을 기록한 기록 매체에 있어서, 비염 중증도 정보, 환경 정보 및 날씨 정보를 포함하는 환자의 특징 정보로부터 상관 분석을 통해 유의한 변수를 추출하여 예측 변수로 결정하는 변수 결정 기능, 예측 변수를 기반으로 비염 진단을 위한 예측 모델을 생성하고, 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정하는 모델 결정 기능 및 예측 변수에 해당되는 환자의 특징 정보를 최적화된 예측 모델에 입력하여 환자의 비염 점수를 예측하는 점수 예측 기능을 구현하는 프로그램이 기록되고 컴퓨터로 읽을 수 있는 기록매체를 제공한다.In another aspect, this embodiment extracts significant variables through correlation analysis from patient characteristic information including rhinitis severity information, environmental information, and weather information in a recording medium recording a program for executing a method for diagnosing rhinitis. Patients corresponding to predictor variables and model determination function that generates a predictive model for rhinitis diagnosis based on the predictor variable, determines the predictive performance of the predictive model, and determines the optimized predictive model. Provides a computer-readable recording medium on which a program implementing a score prediction function for predicting the patient's rhinitis score by inputting feature information of the optimized prediction model is recorded.
도 1은 본 개시가 적용될 수 있는 시스템 구성을 예시적으로 도시한 도면이다. 1 is a diagram exemplarily illustrating a system configuration to which the present disclosure may be applied.
도 2는 본 개시의 일 실시예에 따른 비염 진단 장치의 구성을 도시한 도면이다. 2 is a diagram showing the configuration of a rhinitis diagnosis device according to an embodiment of the present disclosure.
도 3은 본 개시의 일 실시예에 따른 비염 진단 장치가 비염 점수를 예측하는 동작을 설명하기 위한 흐름도이다.Figure 3 is a flow chart for explaining the operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure to predict rhinitis score.
도 4는 본 개시의 일 실시예에 따른 비염 진단 장치가 예측 변수를 결정하는 동작을 설명하기 위한 도면이다. Figure 4 is a diagram for explaining the operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure determines the predictive variable.
도 5는 본 개시의 다른 실시예에 따른 비염 진단 장치가 예측 변수를 결정하는 동작을 설명하기 위한 도면이다. Figure 5 is a diagram for explaining the operation of rhinitis diagnosis apparatus according to another embodiment of the present disclosure determines the predictive variable.
도 6은 본 개시의 일 실시예에 따른 비염 진단 장치가 날씨 정보의 기울기를 이용하는 동작을 설명하기 위한 도면이다.6 is a diagram for explaining an operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure using a slope of weather information.
도 7은 본 개시의 일 실시예에 따른 비염 진단 장치에서 환자의 특징 정보를 입력하는 예시를 도시한 도면이다.7 is a diagram showing an example of inputting characteristic information of a patient in the rhinitis diagnosis apparatus according to an embodiment of the present disclosure.
도 8은 본 개시의 일 실시예에 따른 비염 진단 장치에서 날씨 정보를 이용하는 예시를 도시한 도면이다.8 is a diagram showing an example of using weather information in the rhinitis diagnosis device according to an embodiment of the present disclosure.
도 9은 본 개시의 일 실시예에 따른 비염 진단 장치에서 비염 점수를 예측하는 예시를 도시한 도면이다.9 is a diagram showing an example of predicting a rhinitis score in the rhinitis diagnostic device according to an embodiment of the present disclosure.
도 10은 본 개시의 일 실시예에 따른 비염 진단 방법의 흐름도이다.10 is a flow chart of a method for diagnosing rhinitis according to an embodiment of the present disclosure.
도 11은 본 개시의 일 실시예에 따른 기록매체의 구성을 개념적으로 도시한 도면이다.11 is a diagram conceptually illustrating the configuration of a recording medium according to an embodiment of the present disclosure.
본 개시는 비염 진단 장치, 방법 및 기록매체에 관한 것이다. The present disclosure relates to a device for diagnosing rhinitis, a method, and a recording medium.
이하, 본 개시의 일부 실시예들을 예시적인 도면을 통해 상세하게 설명한다. 각 도면의 구성요소들에 참조부호를 부가함에 있어서, 동일한 구성요소들에 대해서는 비록 다른 도면상에 표시되더라도 가능한 한 동일한 부호를 가지도록 하고 있음에 유의해야 한다. 또한, 본 개시를 설명함에 있어, 관련된 공지 구성 또는 기능에 대한 구체적인 설명이 본 개시의 요지를 흐릴 수 있다고 판단되는 경우에는 그 상세한 설명은 생략한다.Hereinafter, some embodiments of the present disclosure will be described in detail through exemplary drawings. In adding reference numerals to components of each drawing, it should be noted that the same components have the same numerals as much as possible even if they are displayed on different drawings. In addition, in describing the present disclosure, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present disclosure, the detailed description will be omitted.
또한, 본 개시의 구성 요소를 설명하는 데 있어서, 제 1, 제 2, A, B, (a), (b) 등의 용어를 사용할 수 있다. 이러한 용어는 그 구성 요소를 다른 구성 요소와 구별하기 위한 것일 뿐, 그 용어에 의해 해당 구성 요소의 본질이나 차례 또는 순서 등이 한정되지 않는다. 어떤 구성 요소가 다른 구성요소에 "연결", "결합" 또는 "접속"된다고 기재된 경우, 그 구성 요소는 그 다른 구성요소에 직접적으로 연결되거나 또는 접속될 수 있지만, 각 구성 요소 사이에 또 다른 구성 요소가 "연결", "결합" 또는 "접속"될 수도 있다고 이해되어야 할 것이다.Also, terms such as first, second, A, B, (a), and (b) may be used in describing the components of the present disclosure. These terms are only used to distinguish the component from other components, and the nature, order, or order of the corresponding component is not limited by the term. When an element is described as being “connected,” “coupled to,” or “connected” to another element, that element is directly connected or connectable to the other element, but there is another element between the elements. It will be understood that elements may be “connected”, “coupled” or “connected”.
본 명세서에서의 VAS(Visual Analogue Scale) 증상 정보는 환자가 스스로 코 증상에 대해서 중증도에 대해 자가 진단하여 정도를 체크한 시각 상사 점수를 의미할 수 있고, SNOT-22(Sino-Nasal Outcome Test) 정보는 병원에서 검사한 22-항목 비부비동 결과 시험 점수의 평균값을 의미할 수 있다. VAS (Visual Analogue Scale) symptom information in the present specification may mean a visual analog score in which the patient self-diagnoses the severity of nasal symptoms and checks the degree, and SNOT-22 (Sino-Nasal Outcome Test) information may mean the average value of the 22-item nasal sinus outcome test scores tested in the hospital.
본 명세서에서의 비염 점수는 환자의 증상 점수를 기반으로 예측 모델을 통해 예측된 결과로, 병원에서의 비염 진단 관련 변수와 일관성을 가지는 점수를 의미할 수 있다. 또한, 본 명세서에서의 비염 점수는 VAS 코 점수(이지숨)와 동일한 의미로 사용될 수 있다.Rhinitis score in the present specification is a result predicted through a predictive model based on a patient's symptom score, and may mean a score consistent with rhinitis diagnosis-related variables in a hospital. In addition, the rhinitis score in the present specification may be used in the same sense as the VAS nose score (Easysum).
이하 첨부된 도면을 참고하여 본 개시를 상세히 설명하기로 한다.Hereinafter, the present disclosure will be described in detail with reference to the accompanying drawings.
도 1은 본 개시가 적용될 수 있는 시스템 구성을 예시적으로 도시한 도면이다. 1 is a diagram exemplarily illustrating a system configuration to which the present disclosure may be applied.
도 1을 참조하면, 본 개시는 비염 진단 장치을 제공하는 시스템에 관한 것으로, 비염 진단 장치(100) 및 서버(110)에 구현될 수 있다. Referring to Figure 1, the present disclosure relates to a system for providing a rhinitis diagnosis device, may be implemented in the rhinitis diagnosis device 100 and the server (110).
비염 진단 장치(100)는, 일반적인 데스크 탑이나 노트북 등의 일반 PC를 포함하고, 스마트 폰, 태블릿 PC, PDA(Personal Digital Assistants) 및 이동통신 단말기 등의 모바일 단말기 등을 포함할 수 있으며, 이에 제한되지 않고, 서버(110)와 통신 가능한 어떠한 전자 기기로 폭넓게 해석되어야 할 것이다. Rhinitis diagnosis device 100 may include a general PC such as a general desktop or laptop, and may include a mobile terminal such as a smart phone, a tablet PC, a PDA (Personal Digital Assistants) and a mobile communication terminal, and is limited thereto. It is not, and should be broadly interpreted as any electronic device capable of communicating with the server 110.
서버(110)는 하드웨어적으로는 통상적인 웹 서버(Web Server) 또는 웹 어플리케이션 서버(Web Application Server) 또는 웹 서버(WAP Server)와 동일한 구성을 하고 있다. 그러나, 소프트웨어적으로는, 아래에서 상세하게 설명할 바와 같이, C, C++, Java, PHP, .Net, Python, Ruby 등 여하한 언어를 통하여 구현되어 여러 가지 기능을 하는 프로그램 모듈(Module)을 포함할 수 있다.The server 110 has the same configuration as a conventional web server (Web Server), web application server (Web Application Server), or web server (WAP Server) in terms of hardware. However, in terms of software, as will be described in detail below, it includes program modules that are implemented through any language such as C, C++, Java, PHP, .Net, Python, Ruby, and perform various functions. can do.
또한, 서버(110)는 네트워크를 통하여 불특정 다수 클라이언트(비염 진단 장치(100)를 포함) 및/또는 다른 서버와 연결될 수 있는데, 이에 따라, 서버(110)는 클라이언트 또는 다른 서버의 작업수행 요청을 접수하고 그에 대한 작업 결과를 도출하여 제공하는 컴퓨터 시스템 또는 이러한 컴퓨터 시스템을 위하여 설치되어 있는 컴퓨터 소프트웨어(서버 프로그램)를 뜻하는 것일 수도 있다. In addition, the server 110 may be connected to an unspecified number of clients (including the rhinitis diagnosis device 100) and/or other servers through a network. Accordingly, the server 110 receives requests from clients or other servers to perform tasks. It may mean a computer system that accepts and derives and provides work results for it, or computer software (server program) installed for such a computer system.
또한, 서버(110)는 전술한 서버 프로그램 이외에도, 서버(110) 상에서 동작하는 일련의 응용 프로그램(Application Program)과, 경우에 따라서는 내부 또는 외부에 구축되어 있는 각종 데이터베이스를 포함하는 넓은 개념으로 이해되어야 할 것이다. In addition, the server 110 is understood as a broad concept including, in addition to the above-described server program, a series of application programs that operate on the server 110 and, in some cases, various databases built inside or outside. It should be.
여기서, 데이터베이스는, 서버 또는 다른 장치 등에 의해 사용될 목적으로 정보나 자료 등의 데이터가 구조화되어 관리되는 데이터의 집합체를 의미할 수 있으며, 이러한 데이터의 집합체를 저장하는 저장매체를 의미할 수도 있다. Here, the database may refer to an aggregate of data in which data such as information or data is structured and managed for use by a server or other device, and may also refer to a storage medium for storing such an aggregate of data.
또한, 이러한 데이터베이스는 데이터의 구조화 방식, 관리 방식, 종류 등에 따라 분류된 복수의 데이터베이스를 포함하는 것일 수도 있다. 경우에 따라서, 데이터베이스는 정보나 자료 등을 추가, 수정, 삭제 등을 할 수 있도록 해주는 소프트웨어인 데이터베이스 관리시스템(Database Management System, DBMS)을 포함할 수도 있다. In addition, such a database may include a plurality of databases classified according to a data structure method, management method, type, and the like. In some cases, the database may include a database management system (DBMS), which is software that allows information or data to be added, corrected, or deleted.
또한, 서버(110)는 콘텐츠, 각종 정보 및 데이터를 데이터베이스에 저장시키고 관리할 수 있다. 여기서, 데이터베이스는 서버(110)의 내부 또는 외부에 구현될 수 있다.In addition, the server 110 may store and manage contents and various types of information and data in a database. Here, the database may be implemented inside or outside the server 110 .
또한, 서버(110)는 일반적인 서버용 하드웨어에 도스(DOS), 윈도우(windows), 리눅스(Linux), 유닉스(UNIX), 매킨토시(Macintosh) 등의 운영체제에 따라 다양하게 제공되고 있는 서버 프로그램을 이용하여 구현될 수 있으며, 대표적인 것으로는 윈도우 환경에서 사용되는 웹 사이트(Website), IIS(Internet Information Server)와 유닉스환경에서 사용되는 Apache, Nginx, Light HTTP 등이 이용될 수 있다. In addition, the server 110 uses server programs that are provided in various ways according to operating systems such as DOS, Windows, Linux, UNIX, and Macintosh in general server hardware. It can be implemented, and as a representative example, a website, IIS (Internet Information Server) used in a Windows environment, and Apache, Nginx, Light HTTP, etc. used in a Unix environment can be used.
한편, 네트워크(120)는 서버(110)와 비염 진단 장치(100)를 연결해주는 망(Network)으로서, LAN(Local Area Network), WAN(Wide Area Network)등의 폐쇄형 네트워크일 수도 있으나, 인터넷(Internet)과 같은 개방형 네트워크일 수도 있다. 여기서, 인터넷은 TCP/IP 프로토콜 및 그 상위계층에 존재하는 여러 서비스, 즉 HTTP(HyperText Transfer Protocol), Telnet, FTP(File Transfer Protocol), DNS(Domain Name System), SMTP(Simple Mail Transfer Protocol), SNMP(Simple Network Management Protocol), NFS(Network File Service), NIS(Network Information Service)를 제공하는 전 세계적인 개방형 컴퓨터 네트워크 구조를 의미한다. On the other hand, the network 120 is a network that connects the server 110 and the rhinitis diagnosis device 100, and may be a closed network such as a LAN (Local Area Network) and a WAN (Wide Area Network), but the Internet It may be an open network such as (Internet). Here, the Internet refers to the TCP/IP protocol and various services existing in its upper layer, namely HTTP (HyperText Transfer Protocol), Telnet, FTP (File Transfer Protocol), DNS (Domain Name System), SMTP (Simple Mail Transfer Protocol), It refers to a worldwide open computer network structure that provides Simple Network Management Protocol (SNMP), Network File Service (NFS), and Network Information Service (NIS).
또한, 비염 진단 장치(100)가 스마트 폰, 태블릿 PC, PDA(Personal Digital Assistants) 및 이동통신 단말기 등의 모바일 단말기를 포함하는 경우, 네트워크는 이동 통신망이나 와이파이(WiFi) 망 등의 무선 액세스 망을 더 포함할 수도 있다. In addition, when the rhinitis diagnosis device 100 includes a mobile terminal such as a smart phone, a tablet PC, a PDA (Personal Digital Assistants) and a mobile communication terminal, the network is a wireless access network such as a mobile communication network or a Wi-Fi (WiFi) network. may include more.
간략하게 전술한 본 개시의 일 실시예에 따른 비염 진단 장치, 방법 및 기록매체에 대하여, 이하에서 더욱 상세하게 설명한다.With respect to the rhinitis diagnosis device, method and recording medium according to an embodiment of the present disclosure briefly described above, it will be described in more detail below.
본 명세서에서의 비염 진단 장치(100)는 환자의 비염을 진단하기 위해 환자의 특징 정보를 제공받아 비염 점수를 예측하는 유저(User)의 단말을 의미할 수 있다. 또한, 비염 진단 장치(100)는 서버(110)에 의해서 결정되는 예측 모델에 접속할 수 있는 권한을 부여 받은 사용자가 사용하는 장치를 의미할 수 있다. 또한, 비염 진단 장치(100)는 후술한 환자의 특징 정보를 획득하고, 서버로 전송하는 사용자 장치를 의미할 수 있다. 또한, 본 명세서에서의 비염 진단 장치(100)는 서버(110)에 접속하여 컨텐츠 플랫폼에 비염 진단을 위한 정보를 업로드/다운로드 할 수 있다. 여기서, 컨텐츠 플랫폼은 서버(110)에 의해서 운영 또는 동작되는 비염 점수를 예측할 수 있는 온라인 플랫폼을 의미할 수도 있다. Rhinitis diagnosis apparatus 100 in the present specification may refer to a terminal of a user (User) that predicts a rhinitis score by receiving characteristic information of a patient in order to diagnose rhinitis of a patient. In addition, the rhinitis diagnosis device 100 may refer to a device used by a user who has been granted permission to access the predictive model determined by the server 110. In addition, the rhinitis diagnosis apparatus 100 may refer to a user device that obtains the patient's characteristic information described later and transmits it to the server. In addition, the rhinitis diagnosis device 100 in the present specification may connect to the server 110 to upload / download information for diagnosing rhinitis to the content platform. Here, the content platform may mean an online platform capable of predicting rhinitis scores operated or operated by the server 110 .
도 2는 본 개시의 일 실시예에 따른 비염 진단 장치의 구성을 도시한 도면이다. 2 is a diagram showing the configuration of a rhinitis diagnosis device according to an embodiment of the present disclosure.
도 2를 참조하면, 본 개시의 일 실시예에 따른 비염 진단 방법을 제공하는 비염 진단 장치(100)는, 비염 중증도 정보, 환경 정보 및 날씨 정보를 포함하는 환자의 특징 정보로부터 상관 분석을 통해 유의한 변수를 추출하여 예측 변수로 결정하는 변수 결정부(210), 예측 변수를 기반으로 비염 진단을 위한 예측 모델을 생성하고, 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정하는 모델 결정부(220) 및 예측 변수에 해당되는 환자의 특징 정보를 최적화된 예측 모델에 입력하여 환자의 비염 점수를 예측하는 점수 예측부(230)를 포함하는 비염 진단 장치(100)를 제공한다. Referring to FIG. 2, the rhinitis diagnosis apparatus 100, which provides a method for diagnosing rhinitis according to an embodiment of the present disclosure, is significant through correlation analysis from patient characteristic information including rhinitis severity information, environmental information, and weather information. A variable determining unit 210 that extracts a variable and determines it as a predictor variable, a model determining unit that generates a predictive model for diagnosing rhinitis based on the predictor variable, and determines the predictive performance of the predictive model to determine an optimized predictive model. Provides a rhinitis diagnosis device 100 including a score prediction unit 230 for predicting the patient's rhinitis score by inputting the patient's characteristic information corresponding to 220 and the predictor to the optimized predictive model.
일 실시예에 따라 변수 결정부(210)는 비염 중증도 정보, 환경 정보 및 날씨 정보를 포함하는 환자의 특징 정보로부터 상관 분석을 통해 유의한 변수를 추출하여 예측 변수로 결정할 수 있다. 여기서, 비염 중증도 정보는 SNOT-22(Sino-Nasal Outcome Test) 정보, 일상 생활의 지장 유무에 대한 정보 및 VAS 증상 정보 중 적어도 하나의 정보가 포함될 수 있다. 또한, 환경 정보는 곰팡이, 교통 수단, 스트레스 또는 찬공기 등에 관한 정보가 포함될 수 있다.According to an embodiment, the variable determination unit 210 may extract significant variables through correlation analysis from the patient's characteristic information including rhinitis severity information, environmental information, and weather information, and determine them as predictive variables. Here, the rhinitis severity information may include at least one of SNOT-22 (Sino-Nasal Outcome Test) information, information on the presence or absence of disturbances in daily life, and VAS symptom information. In addition, environmental information may include information on mold, transportation, stress, or cold air.
일 예로, 변수 결정부(210)는 환자로부터 입력 인터페이스를 통해서 비염 중증도 정보와 환경 정보를 포함하는 환자의 특징 정보를 입력 받을 수 있다. 여기서, 입력 인터페이스는 터치 스크린, 마이크, 버튼 등 비염 진단 장치(100)에 정보를 입력할 수 있는 모듈 또는 장치를 의미할 수 있다. 다만, 이에 한정되지는 않는다. 예를 들어, 변수 결정부(210)는 환자로부터 입력된 코 증상에 대한 자가 진단 정보인 VAS 증상 정보를 비염 중증도 정보로 획득할 수 있다. 다른 예를 들어, 변수 결정부(210)는 환자로부터 입력된 증상 악화 요인에 관한 정보를 환경 정보로 획득할 수 있다. For example, the variable determining unit 210 may receive input of patient characteristic information including rhinitis severity information and environmental information from the patient through an input interface. Here, the input interface may refer to a module or device capable of inputting information to the rhinitis diagnosis device 100, such as a touch screen, a microphone, and a button. However, it is not limited thereto. For example, the variable determiner 210 may obtain VAS symptom information, which is self-diagnosis information about nasal symptoms input from a patient, as rhinitis severity information. For another example, the variable determiner 210 may obtain information on a symptom aggravating factor input from the patient as environmental information.
다른 일 예로, 변수 결정부(210)는 서버로부터 병원에서 진단받은 환자의 SNOT-22 정보를 수신하여 비염 중증도 정보로 획득할 수 있다. 또한, 변수 결정부(210)는 서버로부터 환자의 위치 정보에 따른 날씨 정보를 수신하여 획득할 수 있다. 예를 들어, 변수 결정부(210)는 위성항법장치(GPS)를 이용한 현재 위치의 온도 정보, 습도 정보 및 미세먼지 정보를 수신하여 날씨 정보로 획득할 수 있다.As another example, the variable determining unit 210 may receive SNOT-22 information of a patient diagnosed in a hospital from a server and obtain it as rhinitis severity information. In addition, the variable determiner 210 may receive and obtain weather information according to the location information of the patient from the server. For example, the variable determiner 210 may receive temperature information, humidity information, and fine dust information of a current location using a global positioning system (GPS) to obtain weather information.
또한, 일 예로, 변수 결정부(210)는 환자의 비염 점수와의 상관 분석을 통해 추출한 유의한 변수를 공통 변수 또는 개별 변수로 구분하여 예측 모델에 적합한 예측 변수를 결정할 수 있다. 예를 들어, 변수 결정부(210)는 환자의 특징 정보의 각 변수에 대한 상관 계수를 산출하고, 상관 계수에 기초하여 추출한 유의한 변수를 공통 변수로 구분할 수 있다. 이 때, 공통 변수는 고정 효과를 가지는 변수를 의미할 수 있다. 구체적으로, 변수 결정부(210)는 환자의 비염 점수와의 상관 분석(Pearson Correlation analysis)을 통해 상관 계수(p)를 산출할 수 있다. 그리고, 변수 결정부(210)는 산출된 상관 계수(p)가 0.3 이상 0.7 이하로 뚜렷한 양적 선형 관계를 갖는 변수 중에서 유의한 변수를 추출하여 공통 변수로 구분할 수 있다. 다른 예를 들어, 변수 결정부(210)는 환자의 위치 정보에 기초하여 획득한 날씨 정보에 포함된 각각의 온도 정보, 습도 정보 및 미세먼지 정보를 개별 변수로 구분할 수 있다. 이 때, 개별 변수는 임의 효과를 가지는 변수를 의미할 수 있다. 또한, 비염 점수에 대한 각각의 온도 정보, 습도 정보 및 미세먼지 정보의 기울기는 환자 별로 상이할 수 있다. 또 다른 예를 들어, 변수 결정부(210)는 임의 절편(random intercept)을 개별 변수로 구분할 수 있다. 구체적으로, 변수 결정부(210)는 비염 점수의 분포가 환자 간에 차이를 보임에 따라 상관 관계를 반영하여 환자 별로 기본 특성을 고려하기 위해 임의 절편을 개별 변수로 구분할 수 있다. 이는 예측되는 비염 점수가 개인마다 다른 주관적인 특성을 가지고 있기 때문이다.In addition, as an example, the variable determining unit 210 may determine a predictor variable suitable for the predictive model by dividing the significant variable extracted through correlation analysis with the rhinitis score of the patient into a common variable or an individual variable. For example, the variable determiner 210 may calculate a correlation coefficient for each variable of the patient's characteristic information, and classify significant variables extracted based on the correlation coefficient as common variables. In this case, the common variable may mean a variable having a fixed effect. Specifically, the variable determination unit 210 may calculate a correlation coefficient (p) through correlation analysis (Pearson Correlation analysis) with the rhinitis score of the patient. Also, the variable determining unit 210 may extract significant variables from variables having a distinct quantitative linear relationship such that the calculated correlation coefficient (p) is 0.3 or more and 0.7 or less, and classify them as common variables. For another example, the variable determiner 210 may classify each of the temperature information, humidity information, and fine dust information included in the obtained weather information based on the location information of the patient as individual variables. In this case, the individual variable may mean a variable having an arbitrary effect. In addition, the slope of each of the temperature information, humidity information, and fine dust information for the rhinitis score may be different for each patient. As another example, the variable determiner 210 may classify a random intercept as an individual variable. Specifically, the variable determiner 210 may classify arbitrary segments as individual variables in order to consider the basic characteristics of each patient by reflecting the correlation as the distribution of rhinitis scores shows a difference between patients. This is because the predicted rhinitis score has different subjective characteristics for each individual.
일 실시예에 따라 모델 결정부(220)는 예측 변수를 기반으로 비염 진단을 위한 예측 모델을 생성하고, 각 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정할 수 있다. 일 예로, 모델 결정부(220)는 예측 변수를 기반으로 회귀 모델(Regression Model), 선형 혼합 모델(Linear Mixed Model) 및 앙상블 모델(Ensemble Machine Learning Model)을 포함하는 복수의 예측 모델을 생성할 수 있다. 그리고, 모델 결정부(220)는 각 예측 모델의 예측 성능을 판단하여 복수의 예측 모델 중에서 최적화된 예측 모델을 결정 할 수 있다. 예를 들어, 모델 결정부(220)는 임의 효과(개별 변수)를 반영하지 않은 회귀 모델, 임의 절편만을 반영한 선형 혼합 모델 또는 임의 절편과 기울기를 모두 반영한 선형 혼합 모델을 예측 모델로 생성할 수 있다. 또한, 모델 결정부(220)는 랜덤 포레스트(Random Forest)알고리즘 기반의 앙상블 모델 또는 LightGBM 알고리즘 기반의 앙상블 모델을 예측 모델로 생성할 수 있다. 이 때, 각각의 앙상블 모델은 임의 효과(개별 변수)를 반영하지 않은 모델, 임의 절편만을 반영한 모델, 임의 절편과 기울기를 모두 반영한 모델로 구성될 수 있다. 다만, 생성되는 모델은 일 예로, 이에 한정되지는 않는다. According to one embodiment, the model determination unit 220 may generate a predictive model for diagnosing rhinitis based on predictor variables, and determine an optimized predictive model by determining the predictive performance of each predictive model. For example, the model determiner 220 may generate a plurality of predictive models including a regression model, a linear mixed model, and an ensemble machine learning model based on predictor variables. there is. And, the model determination unit 220 may determine the prediction performance of each prediction model and determine an optimized prediction model from among a plurality of prediction models. For example, the model determiner 220 may generate a regression model that does not reflect any effect (individual variable), a linear mixed model that reflects only an arbitrary intercept, or a linear mixed model that reflects both the random intercept and the slope as a prediction model. . In addition, the model determiner 220 may generate an ensemble model based on a random forest algorithm or an ensemble model based on a LightGBM algorithm as a predictive model. In this case, each ensemble model may be composed of a model that does not reflect any effect (individual variable), a model that reflects only an arbitrary intercept, or a model that reflects both the random intercept and the slope. However, the generated model is an example, and is not limited thereto.
다른 일 예로, 모델 결정부(220)는 예측 변수에 해당되는 환자의 특징 정보에 K겹 교차 검증을 적용하여 계산한 평균 제곱근 오차를 기반으로 예측 모델의 예측 성능을 판단할 수 있다. 또한, 모델 결정부(220)는 베이즈 정보 기준(Bayesian information criterion, BIC)을 이용하여 모형 적합도를 판단할 수도 있다. 여기서, 베이즈 정보 기준은 복수의 모델 중에서 모델을 선택하는 기준으로 베이지안 통계량에서 사용되는 수치일 수 있다. 예를 들어, 모델 결정부(220)는 복수의 예측 모델 각각을 K겹 교차 검증을 적용하여 계산한 평균 제곱근 오차와 베이즈 정보 기준을 이용하여 최적화된 예측 모델을 결정할 수 있다. 다른 예를 들어, 모델 결정부(220)는 공통 변수 및 개별 변수를 트리 기반의 머신 러닝 알고리즘에 적용하여 생성된 앙상블 모델을 최적화된 예측 모델로 결정할 수 있다.As another example, the model determiner 220 may determine the predictive performance of the predictive model based on the root mean square error calculated by applying K-fold cross-validation to the patient's feature information corresponding to the predictor variable. In addition, the model determination unit 220 may determine the degree of fitness of the model using Bayesian information criterion (BIC). Here, the Bayes information criterion may be a numerical value used in Bayesian statistics as a criterion for selecting a model from among a plurality of models. For example, the model determiner 220 may determine an optimized prediction model using a root mean square error calculated by applying K-fold cross-validation to each of a plurality of prediction models and a Bayes information criterion. For another example, the model determiner 220 may determine an ensemble model generated by applying a common variable and an individual variable to a tree-based machine learning algorithm as an optimized prediction model.
일 실시예에 따라 점수 예측부(230)는 예측 변수에 해당되는 환자의 특징 정보를 최적화된 예측 모델에 입력하여 환자의 비염 점수를 예측할 수 있다. 예를 들어, 점수 예측부(230)는 비염 진단 관련 변수와 예측된 비염 점수와의 내적 일관성을 신뢰도 분석(Cronbach's alpha)을 통해 확인하여 예측 타당성을 획득할 수 있다. 이 때, 비염 진단 관련 변수는 병원에서 비염 진단을 위해 수행한 SNOT-22 정보에 관한 변수일 수 있다.According to one embodiment, the score prediction unit 230 may predict the patient's rhinitis score by inputting the patient's characteristic information corresponding to the predictor variable into an optimized prediction model. For example, the score prediction unit 230 may obtain predictive validity by confirming the internal consistency of rhinitis diagnosis-related variables and the predicted rhinitis score through reliability analysis (Cronbach's alpha). At this time, variables related to rhinitis diagnosis may be variables related to SNOT-22 information performed for rhinitis diagnosis in a hospital.
도 3은 본 개시의 일 실시예에 따른 비염 진단 장치가 비염 점수를 예측하는 동작을 설명하기 위한 흐름도이다.Figure 3 is a flow chart for explaining the operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure to predict rhinitis score.
도 3을 참조하면, 본 개시의 일 실시예에 따른 비염 진단 장치의 변수 결정부(210)는 상관 분석을 통해 유의한 변수를 추출하여 예측 모델에 적합한 예측 변수를 결정할 수 있다(S310). 일 예로, 변수 결정부(210)는 환자의 비염 점수와의 상관 분석을 통해 추출한 유의한 변수를 공통 변수 또는 개별 변수로 구분하여 예측 모델에 적합한 예측 변수를 결정할 수 있다. 예를 들어, 변수 결정부(210)는 환자의 비염 점수와 상관 분석에서 상관 계수(p)를 산출할 수 있다. 그리고, 변수 결정부(210)는 산출된 상관 계수(p)가 0.3 이상 0.7 이하로 뚜렷한 양적 선형 관계를 갖는 변수 중에서 유의한 변수를 추출할 수 있다. 변수 결정부(210)는 추출된 변수 중에서 선형 혼합 모델에 적용하여 유의한 변수만을 예측 변수로 결정할 수 있다. 구체적인 예를 들면, 상관 분석에서 산출된 상관 계수(p)는 비염 중증도에 포함된 SNOT-22 정보 0.66, VAS 증상 점수 0.45, 일상 생활의 지장 유무에 대한 정보 0.36으로 산출될 수 있다. 또한, 상관 계수(p)는 환경 정보에 포함된 곰팡이 0.31, 교통수단 0.34, 스트레스 0.36, 찬공기 0.42 로 산출될 수 있다. 이에 따라 변수 결정부(210)는 상관 계수에 기초하여 SNOT-22 정보, 찬공기, 스트레스, 교통수단에 관한 정보를 최종적으로 공통 변수로 구분하여 예측 변수로 결정할 수 있다. 다만, 환자의 데이터가 변경되면 상관 계수도 변경됨에 따라 이에 한정되는 것은 아니다.Referring to Figure 3, the variable determining unit 210 of the rhinitis diagnostic device according to an embodiment of the present disclosure may determine a predictor variable suitable for the predictive model by extracting a significant variable through a correlation analysis (S310). For example, the variable determination unit 210 may determine a predictor variable suitable for the predictive model by classifying the significant variable extracted through correlation analysis with the rhinitis score of the patient into a common variable or an individual variable. For example, the variable determination unit 210 may calculate a correlation coefficient (p) in the patient's rhinitis score and correlation analysis. Also, the variable determining unit 210 may extract a significant variable from variables having a distinct quantitative linear relationship such that the calculated correlation coefficient (p) is greater than or equal to 0.3 and less than or equal to 0.7. The variable determination unit 210 may determine only significant variables as predictor variables by applying the linear mixed model among the extracted variables. For a specific example, the correlation coefficient (p) calculated in the correlation analysis can be calculated as SNOT-22 information 0.66, VAS symptom score 0.45, information on the presence or absence of disturbance in daily life 0.36 included in rhinitis severity. In addition, the correlation coefficient (p) can be calculated as mold 0.31, transportation 0.34, stress 0.36, and cold air 0.42 included in the environmental information. Accordingly, the variable determining unit 210 may finally classify the SNOT-22 information, cold air, stress, and transportation information as common variables based on the correlation coefficient and determine them as predictive variables. However, it is not limited thereto as the correlation coefficient is also changed when patient data is changed.
다른 일 예로, 변수 결정부(210)는 날씨 정보에 포함된 각각의 온도 정보, 습도 정보 및 미세먼지 정보를 개별 변수로 구분할 수 있다. 예를 들어, 변수 결정부(210)는 환자 별로 비염 점수에 대한 각각의 온도 정보, 습도 정보 및 미세먼지 정보의 기울기는 상이할 수 있다. 따라서, 날씨 정보는 전체 환자를 고려할 때 상관 관계가 낮지만, 서로 다른 기울기로 인해 상쇄됨에 따라 임의 효과를 가지는 개별 변수로 구분할 수 있다. 그리고, 날씨 정보는 기울기(random slope)로 가정하여 예측 모델에 적용될 수도 있다.As another example, the variable determiner 210 may classify each of the temperature information, humidity information, and fine dust information included in the weather information as individual variables. For example, the variable determining unit 210 may have different slopes of temperature information, humidity information, and fine dust information for rhinitis scores for each patient. Accordingly, the weather information has a low correlation when considering the entire patient, but can be divided into individual variables having random effects as they are offset by different slopes. Also, the weather information may be applied to a prediction model by assuming a random slope.
또 다른 일 예로, 변수 결정부(210)는 임의 절편(random intercept)을 개별 변수로 구분할 수 있다. 예를 들어, 변수 결정부(210)는 상관 관계를 반영하여 환자 별로 기본 특성을 다르게 고려하기 위해 임의 절편을 개별 변수로 구분할 수 있다. 이는 예측되는 비염 점수에 환자의 주관적인 비염 중증도 정보가 반영됨에 따라 개인마다 주관적인 특성을 가질 수 있기 때문이다. 또한, 비염 점수의 분포가 환자 간에 차이를 보임에 따라 환자 별로 특징 정보를 다르게 고려해야 하기 때문이다.As another example, the variable determiner 210 may classify a random intercept as an individual variable. For example, the variable determiner 210 may classify the arbitrary intercept as an individual variable in order to consider the basic characteristics differently for each patient by reflecting the correlation. This is because each individual may have subjective characteristics as the patient's subjective rhinitis severity information is reflected in the predicted rhinitis score. In addition, since the distribution of rhinitis scores differs between patients, it is necessary to consider the characteristic information differently for each patient.
일 실시예에 따른 비염 진단 장치의 모델 결정부(220)는 예측 변수를 기반으로 비염 진단을 위한 복수의 예측 모델을 생성할 수 있다(S320). 일 예로, 모델 결정부(220)는 예측 변수를 기반으로 회귀 모델(Regression Model), 선형 혼합 모델(Linear Mixed Model) 및 앙상블 모델(Ensemble Machine Learning Model)을 포함하는 복수의 예측 모델을 생성할 수 있다. 예를 들어, 모델 결정부(220)는 임의 효과(개별 변수)를 반영하지 않은 회귀 모델(Regression Model)을 생성할 수 있다. 여기서, VAS 증상 정보는 SNOT-22 정보와의 상관 관계가 높음에 따라 다중공선성(Multicollinearity) 문제가 발생되어 유의하지 않아 예측 변수에서 제외될 수 있다. 구체적으로, 모델 결정부(220)는 비염 중증도 정보에 포함된 SNOT-22 정보, 일상 생활의 지장 유무에 대한 정보, 환경 정보에 포함된 곰팡이, 교통 수단, 스트레스, 찬공기에 관한 정보 및 절편(rintercept)을 예측 변수로 하는 회귀 모델을 생성할 수 있다. Model determination unit 220 of the rhinitis diagnosis device according to an embodiment may generate a plurality of predictive models for rhinitis diagnosis based on the predictor variables (S320). For example, the model determiner 220 may generate a plurality of predictive models including a regression model, a linear mixed model, and an ensemble machine learning model based on predictor variables. there is. For example, the model determiner 220 may generate a regression model that does not reflect any effect (individual variable). Here, since the VAS symptom information has a high correlation with the SNOT-22 information, a multicollinearity problem occurs and is insignificant, so it can be excluded from the predictor variables. Specifically, the model determination unit 220 is SNOT-22 information included in rhinitis severity information, information on the presence or absence of disturbances in daily life, information on mold, transportation, stress, cold air included in environmental information, and intercepts ( rintercept) as a predictor variable.
다른 예를 들어, 모델 결정부(220)는 임의 절편만을 반영한 선형 혼합 모델 또는 임의 절편과 기울기를 모두 반영한 선형 혼합 모델을 생성할 수 있다. 예를 들어, 모델 결정부(220)는 예측하는 비염 점수가 환자의 주관이 반영되는 특성을 고려하여 선형 혼합 모델을 생성할 수 있다. 여기서, 선형 혼합 모델은 회귀 모델보다 변수의 유의성은 떨어지지만, 환자 내부의 상관 관계를 반영하여 환자 별로 다름을 가정하는 접근으로 합리적인 모델일 수 있다. 구체적인 예를 들면, 모델 결정부(220)는 환자 그룹간 비염 점수가 상이함에 따라 임의 절편만을 반영한 선형 혼합 모델을 생성할 수 있다. 이 때, 예측하는 비염 점수와 상관 관계가 있는 변수는 주로 환경 정보이고, 이 또한 환자 별로 달라지는 주관적인 요인으로 볼 수 있다. 다만, 환경 정보에 포함된 곰팡이에 관한 정보는 해당 선형 혼합 모델에서 유의성이 떨어지고, 환자 중 소수만 악화 요인으로 곰팡이로 입력한 것을 고려하여 예측 변수에서 제외될 수 있다. 따라서, 모델 결정부(220)는 비염 중증도 정보에 포함된 SNOT-22 정보, 환경 정보에 포함된 교통 수단, 스트레스, 찬공기에 관한 정보, 절편(rintercept)을 예측 변수로 하는 회귀 모델을 생성할 수 있다. 구체적인 다른 예를 들면, 모델 결정부(220)는 환자 별로 비염 점수에 대한 날씨 정보의 기울기가 상이함에 따라 임의 절편과 기울기(random slope)를 모두 반영한 선형 혼합 모델을 생성할 수 있다. For another example, the model determiner 220 may generate a linear mixed model reflecting only the random intercept or a linear mixed model reflecting both the random intercept and the slope. For example, the model determining unit 220 may create a linear mixed model in consideration of the characteristic that the predictive rhinitis score reflects the subjectivity of the patient. Here, the linear mixed model is less significant than the regression model, but may be a reasonable model as an approach that assumes that each patient is different by reflecting the correlation within the patient. As a specific example, the model determination unit 220 may generate a linear mixed model reflecting only an arbitrary intercept according to the rhinitis scores of different patient groups. At this time, the variable correlated with the predicted rhinitis score is mainly environmental information, which can also be seen as a subjective factor that varies for each patient. However, the information on fungi included in the environmental information is less significant in the corresponding linear mixed model, and considering that only a small number of patients entered as fungi as an aggravating factor, it can be excluded from the predictor variables. Therefore, the model determination unit 220 generates a regression model using the SNOT-22 information included in the rhinitis severity information, the information on transportation, stress, and cold air included in the environmental information, and the intercept as a predictor variable. can For another specific example, the model determination unit 220 may generate a linear mixed model reflecting both a random intercept and a slope as the slope of the weather information for the rhinitis score is different for each patient.
또 다른 예를 들어, 모델 결정부(220)는 랜덤 포레스트(Random Forest)알고리즘 기반의 앙상블 모델 또는 LightGBM 알고리즘 기반의 앙상블 모델을 생성할 수 있다. 이 때, 각각의 앙상블 모델은 임의 효과(개별 변수)를 반영하지 않은 모델, 임의 절편만을 반영한 모델, 임의 절편과 기울기를 모두 반영한 모델로 구성될 수 있다. 구체적으로, 랜덤 포레스트(Random Forest)알고리즘 기반의 앙상블 모델은 훈련 과정에서 구성한 다수의 결정 트리로부터 분류 또는 예측 결과를 취합해서 최종 결과를 출력하는 모델일 수 있다. 또한, LightGBM 알고리즘 기반의 앙상블 모델은 결정 트리의 오차를 수정하기 위해 최대 손실값을 가지는 노드를 중심으로 계속해서 분할하는 'leaf-wise' 방식을 사용하는 모델일 수 있다.For another example, the model determiner 220 may generate an ensemble model based on a random forest algorithm or an ensemble model based on a LightGBM algorithm. In this case, each ensemble model may be composed of a model that does not reflect any effect (individual variable), a model that reflects only an arbitrary intercept, or a model that reflects both the random intercept and the slope. Specifically, an ensemble model based on a random forest algorithm may be a model that outputs a final result by collecting classification or prediction results from a plurality of decision trees constructed in a training process. In addition, the ensemble model based on the LightGBM algorithm may be a model using a 'leaf-wise' method of continuously splitting around a node having a maximum loss value in order to correct an error of a decision tree.
일 실시예에 따른 비염 진단 장치의 모델 결정부(220)는 각 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정할 수 있다(S330). 일 예로, 모델 결정부(220)는 K겹 교차 검증을 적용하여 계산한 평균 제곱근 오차를 기반으로 예측 모델의 예측 성능을 판단할 수 있다. 예를 들어, 모델 결정부(220)는 데이터를 k개로 분할하여 학습 세트(train set)와 검증 세트(validation set)으로 나누어 k번의 예측 성능을 계산할 수 있다. 여기서 Overall RMSE은 각 iteration별로 계산된 K개 평균 제곱 오차들의 제곱근을 취한 값일 수 있다. Overall RMSE을 이용하여 판단한 각각의 예측 모델의 예측 성능은 표 1과 같다. Model determining unit 220 of the rhinitis diagnosis apparatus according to an embodiment may determine the predictive performance of each predictive model to determine the optimized predictive model (S330). For example, the model determiner 220 may determine the predictive performance of the predictive model based on the root mean square error calculated by applying K-fold cross-validation. For example, the model determiner 220 may calculate prediction performance k times by dividing the data into k pieces and dividing them into a training set and a validation set. Here, Overall RMSE may be a value obtained by taking the square root of K mean square errors calculated for each iteration. Table 1 shows the prediction performance of each prediction model judged using the overall RMSE.
Figure PCTKR2022013025-appb-img-000001
Figure PCTKR2022013025-appb-img-000001
표 1과 같이, 예측 모델의 예측 성능은 선형 혼합 모델과 앙상블 모델에서 임의 효과(개별 변수)의 종류에 따라 다름을 확인할 수 있다. 예를 들어, 선형 혼합 모델에서는 임의 효과(개별 변수)를 제외한 모델의 예측 성능이 좋은 것으로 판단될 수 있다. 반면에, 앙상블 모델에서는 임의 절편만이 반영된 모델의 예측 성능이 좋은 것으로 판단될 수 있다.As shown in Table 1, it can be seen that the predictive performance of the predictive model differs depending on the type of random effect (individual variable) in the linear mixed model and the ensemble model. For example, in a linear mixed model, the predictive performance of the model excluding random effects (individual variables) may be judged to be good. On the other hand, in the ensemble model, it can be determined that the predictive performance of the model in which only the random intercept is reflected is good.
다른 일 예로, 모델 결정부(220)는 베이즈 정보 기준(Bayesian information criterion, BIC)을 이용하여 모형 적합도를 판단할 수도 있다. 예를 들어, 모델 결정부(220)는 회귀 모델의 조정된 결정 계수를 이용하여 모형 적합도를 판단할 수 있다. 여기서 조정된 결정 계수는 회귀 모델의 출력에 포함된 R2으로 해당 모델이 과거 데이터를 잘 설명하는지 측정하는 지표일 수 있다. 구체적인 예를 들면, 생성된 회귀 모델은 조정된 결정 계수로 0.525가 계산되며, 이는 비염 점수 변동의 52.5%를 설명하는 것을 의미할 수 있다. 또한, 임의 절편만을 반영한 선형 혼합 모델은 베이즈 정보 기준에 의한 수치가 1826.423으로 계산될 수 있다. 또한, 임의 절편과 기울기를 모두 반영한 선형 혼합 모델은 해당 수치가 1530.467로 계산되어 임의 절편만을 반영한 모델보다 모형 적합도가 증가한 것으로 판단할 수 있다. 베이즈 정보 기준에 의한 수치는 작을수록 적합한 모델로 판단할 수 있다. 다만, 계산된 수치에 한정되지는 않는다.As another example, the model determiner 220 may determine the degree of fitness of the model using Bayesian information criterion (BIC). For example, the model determiner 220 may determine model fit using the adjusted coefficient of determination of the regression model. Here, the adjusted coefficient of determination is R 2 included in the output of the regression model, and may be an indicator for measuring whether the model explains past data well. As a specific example, the resulting regression model calculated an adjusted coefficient of determination of 0.525, which could mean explaining 52.5% of the rhinitis score variance. In addition, the linear mixed model reflecting only the random intercept can be calculated as 1826.423 based on the Bayes information criterion. In addition, the linear mixed model that reflects both the random intercept and the slope is calculated as 1530.467, so it can be judged that the model fit is higher than the model that reflects only the random intercept. The smaller the numerical value based on the Bayesian information criterion, the more suitable the model can be determined. However, it is not limited to the calculated values.
일 실시예에 따른 비염 진단 장치의 점수 예측부(230)는 최적화된 예측 모델을 이용하여 환자의 비염 점수를 예측할 수 있다(S340). 여기서, 최적화된 예측 모델은 공통 변수 및 개별 변수를 트리 기반의 머신 러닝 알고리즘에 적용하여 생성된 앙상블 모델일 수 있다. 예를 들어, 예측된 환자의 비염 점수는 환자의 특징 정보로부터 결정된 변수들과의 일관성이 확인된 점수일 수 있다. 비염 점수의 일관성 신뢰도에 대한 내용은 도 9를 참조하여 후술한다. Score prediction unit 230 of the rhinitis diagnosis device according to an embodiment may predict the patient's rhinitis score using an optimized prediction model (S340). Here, the optimized prediction model may be an ensemble model generated by applying a tree-based machine learning algorithm to common variables and individual variables. For example, the predicted patient's rhinitis score may be a score for which consistency with variables determined from the patient's characteristic information is confirmed. The consistency reliability of the rhinitis score will be described later with reference to FIG. 9 .
도 4는 본 개시의 일 실시예에 따른 비염 진단 장치가 예측 변수를 결정하는 동작을 설명하기 위한 도면이다.Figure 4 is a diagram for explaining the operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure determines the predictive variable.
도 4를 참조하면, 본 개시의 일 실시예에 따른 비염 진단 장치의 변수 결정부(210)는 연속형 변수와의 상관 분석을 설명할 수 있다. 일 예로, 변수 결정부(210)는 환자의 비염 점수와 비염 중증도 정보에 포함된 연속형 변수의 상관 분석을 통해 유의한 변수를 추출할 수 있다. 예를 들어, 변수 결정부(210)는 비염 점수와 SNOT-22 정보와의 상관 분석을 통해 상관 계수를 0.66으로 산출할 수 있다. 또한, 변수 결정부(210)는 비염 점수와 VAS 증상 정보와의 상관 분석을 통해 상관 계수를 0.45로 산출할 수 있다. 그리고, 변수 결정부(210)는 SNOT-22 정보와 VAS 증상 정보와의 상관 분석을 통해 상관 계수를 0.69로 산출할 수 있다. 따라서, 비염 점수는 SNOT-22 정보 및 VAS 증상 정보와 명확한 양적 선형 관계를 가짐을 확인할 수 있다.Referring to Figure 4, the variable determination unit 210 of the rhinitis diagnosis device according to an embodiment of the present disclosure may explain the correlation analysis with the continuous variable. For example, the variable determining unit 210 may extract significant variables through correlation analysis of continuous variables included in the patient's rhinitis score and rhinitis severity information. For example, the variable determining unit 210 may calculate a correlation coefficient of 0.66 through correlation analysis between the rhinitis score and the SNOT-22 information. In addition, the variable determining unit 210 may calculate a correlation coefficient of 0.45 through correlation analysis between the rhinitis score and the VAS symptom information. Also, the variable determiner 210 may calculate a correlation coefficient of 0.69 through correlation analysis between the SNOT-22 information and the VAS symptom information. Therefore, it can be confirmed that rhinitis score has a clear quantitative linear relationship with SNOT-22 information and VAS symptom information.
도 5는 본 개시의 다른 실시예에 따른 비염 진단 장치가 예측 변수를 결정하는 동작을 설명하기 위한 도면이다.Figure 5 is a diagram for explaining the operation of rhinitis diagnosis apparatus according to another embodiment of the present disclosure determines the predictive variable.
도 5를 참조하면, 본 개시의 일 실시예에 따른 비염 진단 장치의 변수 결정부(210)는 이산형 변수와의 상관 분석을 설명할 수 있다. 일 예로, 변수 결정부(210)는 각 이산형 변수에 대해 비염 점수와 상관 분석을 수행하고 박스 플롯(Box Plot)으로 나타낼 수 있다. 여기서 박스 플롯은 데이터 집합의 범위와 중앙값을 빠르게 확인할 수 있도록 수치적 자료를 표현한 그래프일 수 있다. 예를 들어, 변수 결정부(210)는 비염 점수와 환자로부터 입력된 일상 생활의 지장 유무에 대한 정보와의 상관 분석을 통해 상관 계수를 0.36으로 산출할 수 있다. 또한, 변수 결정부(210)는 비염 점수와 환자로부터 입력된 각각의 환경 정보와의 상관 분석을 통해 상관 계수를 산출할 수 있다. 구체적으로, 변수 결정부(210)는 비염 점수와 곰팡이와의 상관 계수를 0.31로 산출할 수 있다. 변수 결정부(210)는 비염 점수와 교통수단과의 상관 계수를 0.34로 산출할 수 있다. 변수 결정부(210)는 비염 점수와 스트레스와의 상관 계수를 0.36로 산출할 수 있다. 또한, 변수 결정부(210)는 비염 점수와 찬공기와의 상관 계수를 0.42로 산출할 수 있다. 따라서, 비염 점수는 증상 악화 요인에 관한 환경 정보와도 명확한 양적 선형 관계를 가짐을 확인할 수 있다.Referring to Figure 5, the variable determination unit 210 of the rhinitis diagnosis device according to an embodiment of the present disclosure may explain the correlation analysis with the discrete variable. For example, the variable determination unit 210 may perform rhinitis score and correlation analysis for each discrete variable and display it as a box plot. Here, the box plot may be a graph expressing numerical data to quickly check the range and median of the data set. For example, the variable determining unit 210 may calculate a correlation coefficient of 0.36 through correlation analysis with rhinitis scores and information about the presence or absence of disturbances in daily life input from the patient. In addition, the variable determining unit 210 may calculate a correlation coefficient through a correlation analysis between the rhinitis score and each environmental information input from the patient. Specifically, the variable determining unit 210 may calculate a correlation coefficient between the rhinitis score and mold as 0.31. The variable determining unit 210 may calculate a correlation coefficient between the rhinitis score and transportation as 0.34. Variable determining unit 210 may calculate the correlation coefficient between rhinitis score and stress as 0.36. In addition, the variable determiner 210 may calculate a correlation coefficient between the rhinitis score and the cold air as 0.42. Therefore, it can be confirmed that the rhinitis score has a clear quantitative linear relationship with environmental information on symptom aggravating factors.
도 6은 본 개시의 일 실시예에 따른 비염 진단 장치가 날씨 정보의 기울기를 이용하는 동작을 설명하기 위한 도면이다.6 is a diagram for explaining an operation of rhinitis diagnosis apparatus according to an embodiment of the present disclosure using a slope of weather information.
도 6을 참조하면, 본 개시의 일 실시예에 따른 비염 진단 장치의 변수 결정부(210)는 환자 별 날씨 정보에 따른 비염 점수의 분포와 기울기를 설명할 수 있다. 일 예로, 변수 결정부(210)는 환자 별로 위성항법장치(GPS)를 이용하여 현재 위치의 날씨 정보를 획득하고 날씨 정보에 따른 비염 점수의 분포와 기울기를 산출할 수 있다. 예를 들어, 변수 결정부(210)는 환자 별로 비염 점수에 대한 온도 정보, 습도 정보 및 미세먼지 정보의 기울기를 각각 산출할 수 있다. 산출된 기울기는 환자 별로 서로 다른 값을 가질 수 있다. 특히, 도 6은 날씨 정보 중 온도 정보에 따른 비염 점수의 분포와 기울기를 나타낸다.Referring to FIG. 6 , the variable determination unit 210 of the rhinitis diagnosis apparatus according to an embodiment of the present disclosure may explain the distribution and slope of rhinitis scores according to weather information for each patient. For example, the variable determining unit 210 may obtain weather information of a current location for each patient using a global navigation system (GPS) and calculate the distribution and slope of rhinitis scores according to the weather information. For example, the variable determining unit 210 may calculate the slope of temperature information, humidity information, and fine dust information for rhinitis scores for each patient. The calculated slope may have different values for each patient. In particular, FIG. 6 shows the distribution and slope of rhinitis scores according to temperature information among weather information.
도 7은 본 개시의 일 실시예에 따른 비염 진단 장치에서 환자의 특징 정보를 입력하는 예시를 도시한 도면이다.7 is a diagram showing an example of inputting characteristic information of a patient in the rhinitis diagnosis apparatus according to an embodiment of the present disclosure.
도 7을 참조하면, 본 개시의 일 실시예에 따른 비염 진단 장치(100)에서 환자로부터 비염 중증도 정보 및 환경 정보를 획득할 수 있다. 일 예로, 비염 진단 장치(100)에서 환자의 VAS 증상 정보(710)를 획득할 수 있다. 예를 들어, VAS 증상 정보(710)는 환자가가 코 증상에 대해서 증상 자가 진단의 정도를 입력하여 획득하는 것으로, 비염 중증도 정보에 포함되는 정보일 수 있다. 그리고, VAS 증상 정보(710)는 코 증상 이외에 천식 증상에 대해서도 추가로 자가 진단의 정보를 입력하여 획득할 수도 있다. 또한, VAS 증상 정보(710)는 환자가 일정 주기로 입력하되, 일정 기간 수집되어 모니터링될 수 있다. 일정 주기는 1일 간격일 수 있으나 이에 한정되지는 않는다.Referring to Figure 7, it is possible to obtain rhinitis severity information and environmental information from the rhinitis diagnosis apparatus 100 according to an embodiment of the present disclosure from the patient. For example, it is possible to obtain VAS symptom information 710 of the patient in the rhinitis diagnosis apparatus 100 . For example, the VAS symptom information 710 is obtained by inputting the degree of symptom self-diagnosis for nasal symptoms by a patient, and may be information included in rhinitis severity information. In addition, the VAS symptom information 710 may be acquired by inputting additional self-diagnosis information about asthma symptoms in addition to nasal symptoms. In addition, the VAS symptom information 710 may be input by the patient at regular intervals, but collected and monitored for a specific period of time. The predetermined period may be a one-day interval, but is not limited thereto.
다른 일 예로, 비염 진단 장치(100)에서 환자의 악화 요인에 관한 정보(720)를 획득할 수 있다. 예를 들어, 악화 요인에 관한 정보(720)는 환자가 코 증상이 악화된 원인으로 생각되는 요인을 선택하여 획득하는 것으로, 환경 정보에 포함되는 정보일 수 있다. 구체적으로 악화 요인에 관한 정보(720)는 감기, 미세먼지, 집먼지, 애완동물, 곰팡이, 꽃가루, 찬공기, 습도, 교통수단, 냄새, 흡연, 스트레스, 위산역류, 운동, 약물, 음식물 등이 있을 수 있다. 다만, 이에 한정되지는 않는다. As another example, it is possible to obtain information 720 on the aggravating factors of the patient in the rhinitis diagnosis apparatus 100. For example, the information 720 on aggravating factors is obtained by selecting and acquiring a factor that is thought to be a cause of exacerbation of nasal symptoms by a patient, and may be information included in environmental information. Specifically, the information 720 on aggravating factors may include cold, fine dust, house dust, pets, mold, pollen, cold air, humidity, transportation, smell, smoking, stress, acid reflux, exercise, medication, and food. can However, it is not limited thereto.
도 8은 본 개시의 일 실시예에 따른 비염 진단 장치에서 날씨 정보를 이용하는 예시를 도시한 도면이다.8 is a diagram showing an example of using weather information in the rhinitis diagnosis device according to an embodiment of the present disclosure.
도 8을 참조하면, 본 개시의 일 실시예에 따른 비염 진단 장치(100)에서 환자의 날씨 정보(810)를 획득할 수 있다. 일 예로, 날씨 정보(810)는 하루 1회 위성항법장치(GPS)에 기초하여 기상청으로부터 획득한 환자 별 온도 정보, 습도 정보, 미세먼지(PM10) 정보일 수 있다. 다만, 이에 한정되지는 않는다. 예를 들어, 날씨 정보(810)는 환자로부터 화면 터치 등 신호가 입력되면, 비염 진단 장치(100)의 위성항법장치(GPS)를 이용하여 감지한 위치 정보에 기초하여 획득할 수 있다. Referring to FIG. 8 , weather information 810 of a patient may be obtained from the rhinitis diagnosis apparatus 100 according to an embodiment of the present disclosure. For example, the weather information 810 may be temperature information, humidity information, and fine dust (PM10) information for each patient obtained from the Korea Meteorological Administration based on a global navigation system (GPS) once a day. However, it is not limited thereto. For example, the weather information 810 can be obtained based on the location information detected using the global navigation system (GPS) of the rhinitis diagnosis apparatus 100 when a signal such as a screen touch is input from the patient.
도 9은 본 개시의 일 실시예에 따른 비염 진단 장치에서 비염 점수를 예측하는 예시를 도시한 도면이다.9 is a diagram showing an example of predicting a rhinitis score in the rhinitis diagnostic device according to an embodiment of the present disclosure.
도 9을 참조하면, 본 개시의 일 실시예에 따른 비염 진단 장치(100)에서 환자의 비염 점수(910)를 예측할 수 있다. 일 예로, 비염 점수(910)는 환자의 특징 정보를 최적화된 예측 모델에 입력하여 출력한 점수일 수 있다. 예를 들어, 비염 점수(910)는 환자의 비염 중증도 정보에 포함된 변수와의 내적 일관성을 신뢰도 분석(Cronbach's alpha)을 통해 확인할 수 있다. 이에 따라 예측된 비염 점수는 비염 진단으로 대체할 수 있는 수치로서의 예측 타당성을 획득할 수 있다. 여기서, 신뢰도 계수는 검사의 내적 일관성 신뢰도를 나타내는 값으로서, 문항들이 동질적인 요소로 구성되어 있는지를 분석할 수 있다. 구체적으로, 신뢰도 계수는 0.7 이상으로 산출될 수 있다. 이에 따라, 비염 점수(910), SNOT-22정보, 일상 생활의 지장 유무에 대한 정보 및 VAS 증상 정보에 해당되는 4개의 변수의 일관성을 신뢰할 수 있음을 확인할 수 있다. 또한, 비염 진단 장치(100)는 예측된 환자의 비염 점수를 활용하여 조절 정도에 따른 자가 관리 방법을 제공해줄 수 있다. Referring to Figure 9, it is possible to predict the patient's rhinitis score 910 in the rhinitis diagnosis apparatus 100 according to an embodiment of the present disclosure. For example, the rhinitis score 910 may be a score output by inputting patient characteristic information into an optimized predictive model. For example, the rhinitis score 910 can confirm internal consistency with variables included in the patient's rhinitis severity information through reliability analysis (Cronbach's alpha). Accordingly, the predicted rhinitis score can obtain predictive validity as a value that can be replaced with rhinitis diagnosis. Here, the reliability coefficient is a value representing the reliability of the internal consistency of the test, and it can be analyzed whether items are composed of homogeneous elements. Specifically, the reliability coefficient may be calculated to be 0.7 or more. Accordingly, it can be confirmed that the consistency of four variables corresponding to rhinitis score 910, SNOT-22 information, information on the presence or absence of disturbances in daily life, and VAS symptom information is reliable. In addition, the rhinitis diagnosis apparatus 100 may provide a self-management method according to the degree of control by utilizing the predicted patient's rhinitis score.
이하에서는 도 1 내지 도 9을 참조하여 설명한 비염 진단 장치가 수행할 수 있는 비염 진단 방법에 대해서 설명한다. 단, 아래에서는 도 1 내지 도 9에서 설명한 일부 실시예 또는 일부 동작에 대한 상세한 설명을 생략할 수 있으나, 이는 설명의 중복을 방지하기 위한 것일 뿐이므로 비염 진단 방법은 전술한 비염 진단 장치를 동일하게 제공할 수 있다.Hereinafter, a rhinitis diagnosis method that can be performed by the rhinitis diagnosis apparatus described with reference to FIGS. 1 to 9 will be described. However, in the following, detailed descriptions of some embodiments or some operations described in FIGS. 1 to 9 may be omitted, but this is only to prevent duplication of description, so the rhinitis diagnosis method is the same as the rhinitis diagnosis device described above. can provide
도 10은 본 개시의 일 실시예에 따른 비염 진단 방법의 흐름도이다.10 is a flow chart of a method for diagnosing rhinitis according to an embodiment of the present disclosure.
도 10을 참조하면, 본 개시의 비염 진단 방법은 예측 변수를 결정하는 변수 결정 단계를 포함할 수 있다(S1010). 일 예로, 비염 진단 장치는 비염 중증도 정보, 환경 정보 및 날씨 정보를 포함하는 환자의 특징 정보로부터 상관 분석을 통해 유의한 변수를 추출하여 예측 변수로 결정할 수 있다. 예를 들어, 비염 진단 장치는 환자로부터 입력 인터페이스를 통해서 비염 중증도 정보와 환경 정보를 포함하는 환자의 특징 정보를 입력받을 수 있다. 또한, 비염 진단 장치는 환자로부터 입력된 코 증상에 대한 자가 진단 정보인 VAS 증상 정보를 비염 중증도 정보로 획득할 수 있다. 다른 예를 들어, 변수 결정부(210)는 환자로부터 입력된 증상 악화 요인에 관한 정보를 환경 정보로 획득할 수 있다. Referring to Figure 10, the method for diagnosing rhinitis of the present disclosure may include a variable determination step of determining predictive variables (S1010). For example, the rhinitis diagnostic device may extract significant variables through correlation analysis from the patient's characteristic information including rhinitis severity information, environmental information, and weather information and determine them as predictive variables. For example, the rhinitis diagnosis apparatus may receive characteristic information of a patient including rhinitis severity information and environmental information from a patient through an input interface. In addition, the rhinitis diagnosis apparatus may acquire VAS symptom information, which is self-diagnosis information about nasal symptoms input from a patient, as rhinitis severity information. For another example, the variable determiner 210 may obtain information on a symptom aggravating factor input from the patient as environmental information.
다른 일 예로, 비염 진단 장치는 서버로부터 병원에서 진단한 환자의 SNOT-22 정보를 수신하여 비염 중증도 정보로 획득할 수 있다. 또한, 비염 진단 장치는 서버로부터 환자의 위치 정보에 따른 날씨 정보를 수신하여 환자의 특징 정보로 획득할 수 있다. 예를 들어, 비염 진단 장치는 위성항법장치(GPS)를 이용한 현재 위치의 온도 정보, 습도 정보 및 미세먼지 정보를 수신하여 날씨 정보로 획득할 수 있다.As another example, the rhinitis diagnosis device may obtain rhinitis severity information by receiving SNOT-22 information of a patient diagnosed in a hospital from a server. In addition, the rhinitis diagnosis apparatus may receive weather information according to the patient's location information from the server and obtain it as the patient's characteristic information. For example, the rhinitis diagnosis device may obtain weather information by receiving temperature information, humidity information, and fine dust information of the current location using a global navigation system (GPS).
또한, 일 예로, 비염 진단 장치는 환자의 비염 점수와의 상관 분석을 통해 추출한 유의한 변수를 공통 변수 또는 개별 변수로 구분하여 예측 모델에 적합한 예측 변수를 결정할 수 있다. 예를 들어, 비염 진단 장치는 환자의 특징 정보의 각 변수에 대한 상관 계수를 산출하고, 산출된 상관 계수에 기초하여 추출한 유의한 변수를 공통 변수로 구분할 수 있다. 구체적으로, 비염 진단 장치는 환자의 비염 점수와 상관 분석(Pearson Correlation analysis)에서 상관 계수(p)를 산출할 수 있다. 그리고, 비염 진단 장치는 산출된 상관 계수(p)가 0.3 이상 0.7 이하로 뚜렷한 양적 선형 관계를 갖는 변수 중에서 유의한 변수를 추출하여 공통 변수로 구분할 수 있다. 다른 예를 들어, 비염 진단 장치는 환자의 위치 정보에 기초하여 획득한 날씨 정보에 포함된 각각의 온도 정보, 습도 정보 및 미세먼지 정보를 개별 변수로 구분할 수 있다. 환자의 위치 정보에 기초하여 획득한 날씨 정보에 포함된 각각의 온도 정보, 습도 정보 및 미세먼지 정보를 개별 변수로 구분할 수 있다. 또 다른 예를 들어, 비염 진단 장치는 임의 절편(random intercept)을 개별 변수로 구분할 수 있다.In addition, as an example, the rhinitis diagnosis apparatus may determine a predictor variable suitable for a predictive model by dividing a significant variable extracted through correlation analysis with a patient's rhinitis score into a common variable or an individual variable. For example, the rhinitis diagnosis device may calculate a correlation coefficient for each variable of the patient's characteristic information, and distinguish a significant variable extracted based on the calculated correlation coefficient as a common variable. Specifically, the rhinitis diagnosis device may calculate a correlation coefficient (p) from the patient's rhinitis score and correlation analysis (Pearson Correlation analysis). Then, the rhinitis diagnostic device can be divided into common variables by extracting significant variables from among variables having a clear quantitative linear relationship with a calculated correlation coefficient (p) of 0.3 or more and 0.7 or less. For another example, the rhinitis diagnosis device may classify each of the temperature information, humidity information, and fine dust information included in the obtained weather information based on the patient's location information as individual variables. Each of the temperature information, humidity information, and fine dust information included in the obtained weather information based on the patient's location information may be classified as individual variables. As another example, a rhinitis diagnostic device may classify a random intercept as an individual variable.
본 개시의 비염 진단 방법은 예측 모델을 결정하는 모델 결정 단계를 포함할 수 있다(S1020). 일 예로, 비염 진단 장치는 예측 변수를 기반으로 비염 진단을 위한 예측 모델을 생성하고, 각 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정할 수 있다. 예를 들어, 비염 진단 장치는 예측 변수를 기반으로 회귀 모델(Regression Model), 선형 혼합 모델(Linear Mixed Model) 및 앙상블 모델(Ensemble Machine Learning Model)을 포함하는 복수의 예측 모델을 생성할 수 있다. 그리고, 비염 진단 장치는 각 예측 모델의 예측 성능을 판단하여 복수의 예측 모델 중에서 최적화된 예측 모델을 결정 할 수 있다. Rhinitis diagnosis method of the present disclosure may include a model determination step of determining a predictive model (S1020). For example, the rhinitis diagnosis device may generate a predictive model for diagnosing rhinitis based on a predictor variable, and determine an optimized predictive model by determining the predictive performance of each predictive model. For example, the rhinitis diagnosis device may generate a plurality of predictive models including a regression model, a linear mixed model, and an ensemble model (Ensemble Machine Learning Model) based on predictor variables. Then, the rhinitis diagnostic device may determine the prediction performance of each predictive model to determine an optimized predictive model from among a plurality of predictive models.
다른 일 예로, 비염 진단 장치는 예측 변수에 해당되는 환자의 특징 정보에 K겹 교차 검증을 적용하여 계산한 평균 제곱근 오차를 기반으로 예측 모델의 예측 성능을 판단할 수 있다. 또한, 비염 진단 장치는 베이즈 정보 기준(Bayesian information criterion, BIC)을 이용하여 모형 적합도를 판단할 수도 있다. 예를 들어, 비염 진단 장치는 복수의 예측 모델 각각을 K겹 교차 검증을 적용하여 계산한 평균 제곱근 오차와 베이즈 정보 기준이 낮은 값을 가지는 예측 모델을 최적화된 예측 모델을 결정할 수 있다. 다른 예를 들어, 비염 진단 장치는 공통 변수 및 개별 변수를 트리 기반의 머신 러닝 알고리즘에 적용하여 생성된 앙상블 모델을 최적화된 예측 모델로 결정할 수 있다.As another example, the rhinitis diagnosis device may determine the predictive performance of the predictive model based on the root mean square error calculated by applying K-fold cross validation to the patient's characteristic information corresponding to the predictor variable. In addition, rhinitis diagnosis device may determine the model fit using the Bayesian information criterion (BIC). For example, the rhinitis diagnosis device may determine a predictive model optimized for a predictive model having a low value of the root mean square error and Bayesian information criterion calculated by applying K-fold cross-validation to each of a plurality of predictive models. For another example, the rhinitis diagnosis device may determine an ensemble model generated by applying a common variable and an individual variable to a tree-based machine learning algorithm as an optimized predictive model.
본 개시의 비염 진단 방법은 비염 점수를 예측하는 점수 예측 단계를 포함할 수 있다(S1030). 일 예로, 비염 진단 장치는 예측 변수에 해당되는 환자의 특징 정보를 최적화된 예측 모델에 입력하여 환자의 비염 점수를 예측할 수 있다. 예를 들어, 비염 진단 장치는 비염 진단 관련 변수와 예측된 비염 점수와의 내적 일관성을 신뢰도 분석(Cronbach's alpha)을 통해 확인하여 예측 타당성을 획득할 수 있다. Rhinitis diagnosis method of the present disclosure may include a score prediction step of predicting the rhinitis score (S1030). For example, the rhinitis diagnostic device may predict the patient's rhinitis score by inputting the patient's characteristic information corresponding to the predictor variable into an optimized predictive model. For example, the rhinitis diagnosis device can obtain predictive validity by confirming the internal consistency between the rhinitis diagnosis-related variables and the predicted rhinitis score through reliability analysis (Cronbach's alpha).
이하에서는 도 10을 참조하여 전술한 비염 진단 방법을 실행시키기 위한 프로그램을 기록한 기록매체에 포함되는 기능에 대해서 설명한다. 단, 아래에서는 도 1 내지 도 9에서 설명한 일부 실시예 또는 일부 동작에 대한 상세한 설명을 생략할 수 있으나, 이는 설명의 중복을 방지하기 위한 것일 뿐이므로 전술한 비염 진단 방법에 대응되는 모든 기능을 실행할 수 있다. Hereinafter, with reference to Figure 10 will be described the functions included in the recording medium recording the program for executing the rhinitis diagnosis method described above. However, in the following, detailed descriptions of some embodiments or some operations described in FIGS. 1 to 9 may be omitted, but this is only to prevent duplication of description, so all functions corresponding to the rhinitis diagnosis method described above can be executed. can
도 11은 본 개시의 일 실시예에 따른 기록매체의 구성을 개념적으로 도시한 도면이다.11 is a diagram conceptually illustrating the configuration of a recording medium according to an embodiment of the present disclosure.
일 실시예에 따른 비염 진단 방법을 실행시키기 위한 프로그램을 기록한 기록매체(1100)는 비염 중증도 정보, 환경 정보 및 날씨 정보를 포함하는 환자의 특징 정보로부터 상관 분석을 통해 유의한 변수를 추출하여 예측 변수로 결정하는 변수 결정 기능(1110), 예측 변수를 기반으로 비염 진단을 위한 예측 모델을 생성하고, 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정하는 모델 결정 기능(1120) 및 예측 변수에 해당되는 환자의 특징 정보를 최적화된 예측 모델에 입력하여 환자의 비염 점수를 예측하는 점수 예측 기능(1130)을 포함할 수 있다. The recording medium 1100 recording the program for executing the method for diagnosing rhinitis according to an embodiment extracts significant variables through correlation analysis from patient characteristic information including rhinitis severity information, environmental information, and weather information, and predicts variables. A variable determining function 1110 that determines a model determining function 1120 that generates a predictive model for diagnosing rhinitis based on the predictor and determines an optimized predictive model by determining the predictive performance of the predictor and predictor It may include a score prediction function 1130 for predicting the patient's rhinitis score by inputting the corresponding patient's characteristic information into an optimized prediction model.
일 예에 따라, 변수 결정 기능(1110)은 비염 중증도 정보, 환경 정보 및 날씨 정보를 포함하는 환자의 특징 정보로부터 상관 분석을 통해 유의한 변수를 추출하여 예측 변수로 결정할 수 있다. 일 예로, 변수 결정 기능(1110)은 환자의 비염 점수와의 상관 분석을 통해 추출한 유의한 변수를 공통 변수 또는 개별 변수로 구분하여 예측 모델에 적합한 예측 변수를 결정할 수 있다. 예를 들어, 변수 결정 기능(1110)은 환자의 특징 정보의 각 변수에 대한 상관 계수를 산출하고, 상관 계수에 기초하여 추출한 유의한 변수를 공통 변수로 구분할 수 있다. 또한, 변수 결정 기능(1110)은 환자의 위치 정보에 기초하여 획득한 날씨 정보에 포함된 각각의 온도 정보, 습도 정보 및 미세먼지 정보를 개별 변수로 구분하며, 비염 점수에 대한 각각의 온도 정보, 습도 정보 및 미세먼지 정보의 기울기는 환자 별로 상이할 수 있다. According to an example, the variable determining function 1110 may extract a significant variable through correlation analysis from the patient's characteristic information including rhinitis severity information, environmental information, and weather information and determine it as a predictor variable. For example, the variable determining function 1110 may determine a predictor variable suitable for a predictive model by classifying a significant variable extracted through correlation analysis with a patient's rhinitis score into a common variable or an individual variable. For example, the variable determination function 1110 may calculate a correlation coefficient for each variable of the patient's characteristic information, and classify significant variables extracted based on the correlation coefficient as common variables. In addition, the variable determination function 1110 classifies each of the temperature information, humidity information, and fine dust information included in the obtained weather information based on the patient's location information as individual variables, each temperature information for the rhinitis score, The slope of the humidity information and the fine dust information may be different for each patient.
일 예에 따라, 모델 결정 기능(1120)은 예측 변수를 기반으로 비염 진단을 위한 예측 모델을 생성하고, 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정할 수 있다. 일 예로, 모델 결정 기능(1120)은 예측 변수를 기반으로 회귀 모델, 선형 혼합 모델 및 앙상블 모델을 포함하는 복수의 예측 모델을 생성할 수 있다. 그리고 모델 결정 기능(1120)은 각 예측 모델의 예측 성능을 판단하여 복수의 예측 모델 중에서 최적화된 예측 모델을 결정할 수 있다. 예를 들어, 모델 결정 기능(1120)은 예측 변수에 해당되는 환자의 특징 정보에 K겹 교차 검증을 적용하여 계산한 평균 제곱근 오차를 기반으로 예측 모델의 예측 성능을 판단할 수 있다. 다른 예를 들어, 모델 결정 기능(1120)은 공통 변수 및 개별 변수를 트리 기반의 머신 러닝 알고리즘에 적용하여 생성된 앙상블 모델을 최적화된 예측 모델로 결정할 수 있다.According to one example, the model determining function 1120 may generate a predictive model for diagnosing rhinitis based on predictor variables and determine an optimized predictive model by determining predictive performance of the predictive model. For example, the model determination function 1120 may generate a plurality of predictive models including a regression model, a linear mixed model, and an ensemble model based on predictor variables. In addition, the model determination function 1120 may determine the prediction performance of each prediction model and determine an optimized prediction model from among a plurality of prediction models. For example, the model determination function 1120 may determine the predictive performance of the predictive model based on the root mean square error calculated by applying K-fold cross-validation to the patient's feature information corresponding to the predictor variable. For another example, the model determination function 1120 may determine an ensemble model generated by applying a common variable and an individual variable to a tree-based machine learning algorithm as an optimized predictive model.
전술한 본 개시의 실시예에 따른 비염 진단 방법은 비염 진단 장치(100)에 기본적으로 설치되거나 사용자에 의해 직접 설치된 애플리케이션(즉, 프로그램)으로 구현되고, 비염 진단 장치(100) 등의 컴퓨터로 읽을 수 있는 기록매체에 기록될 수 있다. The rhinitis diagnosis method according to an embodiment of the present disclosure described above is implemented as an application (ie, a program) installed by default in the rhinitis diagnosis device 100 or directly installed by a user, and is readable by a computer such as the rhinitis diagnosis device 100. can be recorded on a recordable medium.
본 개시의 실시예에 따른 비염 진단 방법을 구현한 프로그램은, 변수 결정 기능, 모델 결정 기능, 점수 예측 기능 등을 실행한다. 이러한 프로그램은 컴퓨터에 의해 읽힐 수 있는 기록매체에 기록되고 컴퓨터에 의해 실행됨으로써 전술한 기능들이 실행될 수 있다. A program implementing the method for diagnosing rhinitis according to an embodiment of the present disclosure executes a variable determination function, a model determination function, a score prediction function, and the like. These programs can be recorded on a computer-readable recording medium and executed by a computer to execute the aforementioned functions.
이와 같이, 컴퓨터가 기록매체에 기록된 프로그램을 읽어 들여 프로그램으로 구현된 비염 진단 방법을 실행시키기 위하여, 전술한 프로그램은 컴퓨터의 프로세서(CPU)가 읽힐 수 있는 C, C++, JAVA, 기계어 등의 컴퓨터 언어로 코드화된 코드(Code)를 포함할 수 있다. In this way, in order for the computer to read the program recorded on the recording medium and execute the rhinitis diagnosis method implemented as a program, the above-described program is a computer such as C, C ++, JAVA, machine language, etc. that the processor (CPU) of the computer can read. It may include code coded in a language.
이러한 코드는 전술한 기능들을 정의한 함수 등과 관련된 기능적인 코드(Function Code)를 포함할 수 있고, 전술한 기능들을 컴퓨터의 프로세서가 소정의 절차대로 실행시키는데 필요한 실행 절차 관련 제어 코드를 포함할 수도 있다. These codes may include functional codes related to functions defining the above-described functions, and may include control codes related to execution procedures necessary for a processor of a computer to execute the above-described functions according to a predetermined procedure.
또한, 이러한 코드는 전술한 기능들을 컴퓨터의 프로세서가 실행시키는데 필요한 추가 정보나 미디어가 컴퓨터의 내부 또는 외부 메모리의 어느 위치(주소 번지)에서 참조 되어야 하는지에 대한 메모리 참조 관련 코드를 더 포함할 수 있다. In addition, these codes may further include memory reference related codes for which location (address address) of the computer's internal or external memory should be referenced for additional information or media necessary for the computer's processor to execute the above-mentioned functions. .
또한, 컴퓨터의 프로세서가 전술한 기능들을 실행시키기 위하여 원격(Remote)에 있는 어떠한 다른 컴퓨터나 서버 등과 통신이 필요한 경우, 코드는 컴퓨터의 프로세서가 컴퓨터의 통신 모듈(예: 유선 및/또는 무선 통신 모듈)을 이용하여 원격(Remote)에 있는 어떠한 다른 컴퓨터나 서버 등과 어떻게 통신해야만 하는지, 통신 시 어떠한 정보나 미디어를 송수신해야 하는지 등에 대한 통신 관련 코드를 더 포함할 수도 있다. In addition, when the computer processor needs to communicate with any other remote computer or server in order to execute the above-mentioned functions, the code is used by the computer processor to communicate with the computer's communication module (e.g., wired and/or wireless communication module). ) may further include communication-related codes for how to communicate with any other remote computer or server, and what information or media should be transmitted/received during communication.
그리고, 본 개시를 구현하기 위한 기능적인(Functional) 프로그램과 이와 관련된 코드 및 코드 세그먼트 등은, 기록매체를 읽어서 프로그램을 실행시키는 컴퓨터의 시스템 환경 등을 고려하여, 본 개시가 속하는 기술분야의 프로그래머들에 의해 용이하게 추론되거나 변경될 수도 있다.In addition, a functional program for implementing the present disclosure, codes and code segments related thereto, in consideration of the system environment of a computer that reads a recording medium and executes a program, etc. It may be easily inferred or changed by
또한 전술한 바와 같은 프로그램을 기록한 컴퓨터로 읽힐 수 있는 기록매체는 네트워크로 커넥션된 컴퓨터 시스템에 분산되어, 분산방식으로 컴퓨터가 읽을 수 있는 코드가 저장되고 실행될 수 있다. 이 경우, 다수의 분산된 컴퓨터 중 어느 하나 이상의 컴퓨터는 상기에 제시된 기능들 중 일부를 실행하고, 그 결과를 다른 분산된 컴퓨터들 중 하나 이상에 그 실행 결과를 전송할 수 있으며, 그 결과를 전송받은 컴퓨터 역시 상기에 제시된 기능들 중 일부를 실행하여, 그 결과를 역시 다른 분산된 컴퓨터들에 제공할 수 있다. In addition, the computer-readable recording medium on which the above-described program is recorded is distributed to computer systems connected through a network, so that computer-readable codes can be stored and executed in a distributed manner. In this case, any one or more of the plurality of distributed computers may execute some of the functions presented above, transmit the execution results to one or more of the other distributed computers, and receive the transmitted results. A computer may also execute some of the functions presented above and provide the results to other distributed computers as well.
이상에서 전술한 바와 같은, 본 개시의 실시예에 따른 비염 진단 방법을 실행시키기 위한 프로그램을 기록한 컴퓨터로 읽힐 수 있는 기록매체는, 일 예로, ROM, RAM, CD-ROM, 자기 테이프, 플로피디스크, 광 미디어 저장장치 등이 있다. As described above, a computer-readable recording medium recording a program for executing the rhinitis diagnosis method according to an embodiment of the present disclosure is, for example, ROM, RAM, CD-ROM, magnetic tape, floppy disk, and optical media storage devices.
또한, 본 개시의 실시예에 따른 비염 진단 방법을 실행시키기 위한 프로그램인 애플리케이션을 기록한 컴퓨터로 읽을 수 있는 기록매체는, 애플리케이션 스토어 서버(Application Store Server), 애플리케이션 또는 해당 서비스와 관련된 웹 서버(Web Server) 등을 포함하는 애플리케이션 제공 서버(Application Provider Server)에 포함된 저장매체(예: 하드디스크 등)이거나, 애플리케이션 제공 서버 그 자체일 수도 있으며, 프로그램을 기록한 다른 컴퓨터 또는 그 저장매체일 수도 있다. In addition, the computer-readable recording medium recording the application, which is a program for executing the rhinitis diagnosis method according to an embodiment of the present disclosure, is an application store server (Application Store Server), an application or a web server related to the service (Web Server ), etc., may be a storage medium (eg, hard disk, etc.) included in the application providing server (Application Provider Server), the application providing server itself, or another computer on which a program is recorded or its storage medium.
본 개시의 일 실시예에 따른 비염 진단 방법을 실행시키기 위한 프로그램인 애플리케이션을 기록한 기록매체를 읽을 수 있는 컴퓨터는, 일반적인 데스크 탑이나 노트북 등의 일반 PC 뿐만 아니라, 스마트 폰, 태블릿 PC, PDA(Personal Digital Assistants) 및 이동통신 단말기 등의 모바일 단말기를 포함할 수 있으며, 이뿐만 아니라, 컴퓨팅(Computing) 가능한 모든 기기로 해석되어야 할 것이다. A computer capable of reading a recording medium on which an application, which is a program for executing the rhinitis diagnosis method according to an embodiment of the present disclosure, is recorded, as well as a general PC such as a general desktop or laptop computer, a smart phone, a tablet PC, a PDA (Personal It may include mobile terminals such as Digital Assistants and mobile communication terminals, and should be interpreted as all devices capable of computing.
만약, 본 개시의 일 실시예에 따른 비염 진단 방법을 실행시키기 위한 프로그램인 애플리케이션을 기록한 기록매체를 읽을 수 있는 컴퓨터가 스마트 폰, 태블릿 PC, PDA(Personal Digital Assistants) 및 이동통신 단말기 등의 모바일 단말기인 경우, 모바일 단말기는 애플리케이션 스토어 서버, 웹 서버 등을 포함하는 애플리케이션 제공 서버로부터 해당 애플리케이션을 다운로드 받아 설치할 수 있고, 경우에 따라서는, 애플리케이션 제공 서버에서 일반 PC로 다운로드 된 이후, 동기화 프로그램을 통해 모바일 단말기에 설치될 수도 있다. If a computer capable of reading a recording medium on which an application, which is a program for executing the rhinitis diagnosis method according to an embodiment of the present disclosure, is recorded, a mobile terminal such as a smart phone, tablet PC, PDA (Personal Digital Assistants) and mobile communication terminal In this case, the mobile terminal may download and install the corresponding application from an application providing server including an application store server, a web server, etc. In some cases, after being downloaded from the application providing server to a general PC, the mobile device is downloaded through a synchronization program. It can also be installed in a terminal.
이상에서, 본 개시의 실시예를 구성하는 모든 구성 요소들이 하나로 결합되거나 결합되어 동작하는 것으로 설명되었다고 해서, 본 개시가 반드시 이러한 실시예에 한정되는 것은 아니다. 즉, 본 개시의 목적 범위 안에서라면, 그 모든 구성 요소들이 하나 이상으로 선택적으로 결합하여 동작할 수도 있다. 또한, 그 모든 구성 요소들이 각각 하나의 독립적인 하드웨어로 구현될 수 있지만, 각 구성 요소들의 그 일부 또는 전부가 선택적으로 조합되어 하나 또는 복수 개의 하드웨어에서 조합된 일부 또는 전부의 기능을 수행하는 프로그램 모듈을 갖는 컴퓨터 프로그램으로서 구현될 수도 있다. 그 컴퓨터 프로그램을 구성하는 코드들 및 코드 세그먼트들은 본 개시의 기술 분야의 당업자에 의해 용이하게 추론될 수 있을 것이다. 이러한 컴퓨터 프로그램은 컴퓨터가 읽을 수 있는 저장매체(Computer Readable Media)에 저장되어 컴퓨터에 의하여 읽혀지고 실행됨으로써, 본 개시의 실시예를 구현할 수 있다. 컴퓨터 프로그램의 저장매체로서는 자기 기록매체, 광 기록매체, 등이 포함될 수 있다.In the above, even though all components constituting the embodiments of the present disclosure have been described as being combined or operated as one, the present disclosure is not necessarily limited to these embodiments. That is, within the scope of the purpose of the present disclosure, all of the components may be selectively combined with one or more to operate. In addition, although all of the components may be implemented as a single independent piece of hardware, some or all of the components are selectively combined to perform some or all of the combined functions in one or a plurality of hardware. It may be implemented as a computer program having. Codes and code segments constituting the computer program may be easily inferred by a person skilled in the art of the present disclosure. Such a computer program may implement an embodiment of the present disclosure by being stored in a computer readable storage medium, read and executed by a computer. A storage medium of a computer program may include a magnetic recording medium, an optical recording medium, and the like.
또한, 이상에서 기재된 "포함하다", "구성하다" 또는 "가지다" 등의 용어는, 특별히 반대되는 기재가 없는 한, 해당 구성 요소가 내재될 수 있음을 의미하는 것이므로, 다른 구성 요소를 제외하는 것이 아니라 다른 구성 요소를 더 포함할 수 있는 것으로 해석되어야 한다. 기술적이거나 과학적인 용어를 포함한 모든 용어들은, 다르게 정의되지 않는 한, 본 개시가 속하는 기술 분야에서 통상의 지식을 가진 자에 의해 일반적으로 이해되는 것과 동일한 의미를 가진다. 사전에 정의된 용어와 같이 일반적으로 사용되는 용어들은 관련 기술의 문맥 상의 의미와 일치하는 것으로 해석되어야 하며, 본 개시에서 명백하게 정의하지 않는 한, 이상적이거나 과도하게 형식적인 의미로 해석되지 않는다.In addition, terms such as "include", "comprise" or "have" described above mean that the corresponding component may be inherent unless otherwise stated, excluding other components. It should be construed as being able to further include other components. All terms, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs, unless defined otherwise. Commonly used terms, such as terms defined in a dictionary, should be interpreted as consistent with the meaning in the context of the related art, and are not interpreted in an ideal or excessively formal meaning unless explicitly defined in the present disclosure.
이상의 설명은 본 개시의 기술 사상을 예시적으로 설명한 것에 불과한 것으로서, 본 개시가 속하는 기술 분야에서 통상의 지식을 가진 자라면 본 개시의 본질적인 특성에서 벗어나지 않는 범위에서 다양한 수정 및 변형이 가능할 것이다. 따라서, 본 개시에 개시된 실시예들은 본 개시의 기술 사상을 한정하기 위한 것이 아니라 설명하기 위한 것이고, 이러한 실시예에 의하여 본 개시의 기술 사상의 범위가 한정되는 것은 아니다. 본 개시의 보호 범위는 아래의 청구범위에 의하여 해석되어야 하며, 그와 동등한 범위 내에 있는 모든 기술 사상은 본 개시의 권리범위에 포함되는 것으로 해석되어야 할 것이다.The above description is merely an example of the technical idea of the present disclosure, and various modifications and variations may be made to those skilled in the art without departing from the essential characteristics of the present disclosure. Therefore, the embodiments disclosed in this disclosure are not intended to limit the technical spirit of the present disclosure, but to explain, and the scope of the technical spirit of the present disclosure is not limited by these embodiments. The protection scope of the present disclosure should be construed by the claims below, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present disclosure.
CROSS-REFERENCE TO RELATED APPLICATIONCROSS-REFERENCE TO RELATED APPLICATION
본 특허출원은 2021년 09월 07일 한국에 출원한 특허출원번호 제 10-2021-0118903호에 대해 미국 특허법 119(a)조 (35 U.S.C § 119(a))에 따라 우선권을 주장하며, 그 모든 내용은 참고문헌으로 본 특허출원에 병합된다. 아울러, 본 특허출원은 미국 이외에 국가에 대해서도 위와 동일한 이유로 우선권을 주장하면 그 모든 내용은 참고문헌으로 본 특허출원에 병합된다.This patent application claims priority in accordance with US Patent Act Article 119 (a) (35 U.S.C § 119 (a)) for Patent Application No. 10-2021-0118903 filed in Korea on September 7, 2021, and All contents are incorporated into this patent application by reference. In addition, if this patent application claims priority for the same reason as above for countries other than the United States, all the contents are incorporated into this patent application as references.

Claims (15)

  1. 비염 중증도 정보, 환경 정보 및 날씨 정보를 포함하는 환자의 특징 정보로부터 상관 분석을 통해 유의한 변수를 추출하여 예측 변수로 결정하는 변수 결정부;Variable determination unit for extracting significant variables from the patient's characteristic information including rhinitis severity information, environmental information and weather information through correlation analysis and determining them as predictive variables;
    상기 예측 변수를 기반으로 비염 진단을 위한 예측 모델을 생성하고, 상기 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정하는 모델 결정부; 및A model determination unit for generating a predictive model for diagnosing rhinitis based on the predictive variables and determining an optimized predictive model by determining predictive performance of the predictive model; and
    상기 예측 변수에 해당되는 상기 환자의 특징 정보를 상기 최적화된 예측 모델에 입력하여 환자의 비염 점수를 예측하는 점수 예측부를 포함하는 것을 특징으로 하는 비염 진단 장치.Rhinitis diagnostic device comprising a score prediction unit for predicting the patient's rhinitis score by inputting the patient's characteristic information corresponding to the predictor variable into the optimized predictive model.
  2. 제 1 항에 있어서,According to claim 1,
    상기 변수 결정부는,The variable determination unit,
    상기 환자의 비염 점수와의 상관 분석을 통해 추출한 유의한 변수를 공통 변수 또는 개별 변수로 구분하여 상기 예측 모델에 적합한 상기 예측 변수를 결정하는 것을 특징으로 하는 비염 진단 장치.Rhinitis diagnostic device, characterized in that for determining the predictor variable suitable for the predictive model by dividing the significant variable extracted through the correlation analysis with the rhinitis score of the patient into a common variable or an individual variable.
  3. 제 2 항에 있어서,According to claim 2,
    상기 변수 결정부는,The variable determination unit,
    상기 환자의 특징 정보의 각 변수에 대한 상관 계수를 산출하고, 상기 상관 계수에 기초하여 추출한 유의한 변수를 상기 공통 변수로 구분하는 것을 특징으로 하는 비염 진단 장치.Rhinitis diagnosis device, characterized in that for calculating the correlation coefficient for each variable of the patient's characteristic information, and dividing the significant variable extracted based on the correlation coefficient as the common variable.
  4. 제 2 항에 있어서,According to claim 2,
    상기 변수 결정부는,The variable determination unit,
    환자의 위치 정보에 기초하여 획득한 상기 날씨 정보에 포함된 각각의 온도 정보, 습도 정보 및 미세먼지 정보를 상기 개별 변수로 구분하며,Each of the temperature information, humidity information, and fine dust information included in the weather information acquired based on the patient's location information is classified as the individual variable,
    상기 비염 점수에 대한 상기 각각의 온도 정보, 습도 정보 및 미세먼지 정보의 기울기는 환자 별로 상이한 것을 특징으로 하는 비염 진단 장치.Rhinitis diagnosis device, characterized in that the slope of each of the temperature information, humidity information and fine dust information for the rhinitis score is different for each patient.
  5. 제 1 항에 있어서,According to claim 1,
    상기 모델 결정부는,The model determining unit,
    상기 예측 변수를 기반으로 회귀 모델, 선형 혼합 모델 및 앙상블 모델을 포함하는 복수의 예측 모델을 생성하고, 각 예측 모델의 예측 성능을 판단하여 상기 복수의 예측 모델 중에서 상기 최적화된 예측 모델을 결정하는 것을 특징으로 하는 비염 진단 장치.Generating a plurality of predictive models including a regression model, a linear mixed model, and an ensemble model based on the predictor variables, and determining the optimized predictive model among the plurality of predictive models by determining the predictive performance of each predictive model. Rhinitis diagnosis device characterized by.
  6. 제 5 항에 있어서,According to claim 5,
    상기 모델 결정부는,The model determining unit,
    상기 예측 변수에 해당되는 상기 환자의 특징 정보에 K겹 교차 검증(K fold Cross Validation)을 적용하여 계산한 평균 제곱근 오차(Root Mean Square Error, RMSE)를 기반으로 상기 예측 모델의 예측 성능을 판단하는 것을 특징으로 하는 비염 진단 장치.Based on the Root Mean Square Error (RMSE) calculated by applying K-fold Cross Validation to the patient's feature information corresponding to the predictor variable, Determining the predictive performance of the predictive model Rhinitis diagnostic device, characterized in that.
  7. 제 5 항에 있어서,According to claim 5,
    상기 최적화된 예측 모델은,The optimized predictive model,
    공통 변수 및 개별 변수를 트리 기반의 머신 러닝 알고리즘에 적용하여 생성된 상기 앙상블 모델인 것을 특징으로 하는 비염 진단 장치.Rhinitis diagnostic device, characterized in that the ensemble model generated by applying a common variable and individual variables to a tree-based machine learning algorithm.
  8. 비염 중증도 정보, 환경 정보 및 날씨 정보를 포함하는 환자의 특징 정보로부터 상관 분석을 통해 유의한 변수를 추출하여 예측 변수로 결정하는 변수 결정 단계;A variable determination step of extracting significant variables from the patient's characteristic information including rhinitis severity information, environmental information, and weather information through correlation analysis and determining them as predictive variables;
    상기 예측 변수를 기반으로 비염 진단을 위한 예측 모델을 생성하고, 상기 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정하는 모델 결정 단계; 및A model determination step of generating a predictive model for diagnosing rhinitis based on the predictive variables and determining an optimized predictive model by determining predictive performance of the predictive model; and
    상기 예측 변수에 해당되는 상기 환자의 특징 정보를 상기 최적화된 예측 모델에 입력하여 환자의 비염 점수를 예측하는 점수 예측 단계를 포함하는 것을 특징으로 하는 비염 진단 방법.Rhinitis diagnosis method characterized in that it comprises a score prediction step of predicting the patient's rhinitis score by inputting the patient's characteristic information corresponding to the predictor variable into the optimized predictive model.
  9. 제 8 항에 있어서,According to claim 8,
    상기 변수 결정 단계는,The variable determination step,
    상기 환자의 비염 점수와의 상관 분석을 통해 추출한 유의한 변수를 공통 변수 또는 개별 변수로 구분하여 상기 예측 모델에 적합한 상기 예측 변수를 결정하는 것을 특징으로 하는 비염 진단 방법.Rhinitis diagnosis method, characterized in that for determining the predictor variable suitable for the predictive model by dividing the significant variable extracted through the correlation analysis with the rhinitis score of the patient into a common variable or an individual variable.
  10. 제 9 항에 있어서,According to claim 9,
    상기 변수 결정 단계는,The variable determination step,
    상기 환자의 특징 정보의 각 변수에 대한 상관 계수를 산출하고, 상기 상관 계수에 기초하여 추출한 유의한 변수를 상기 공통 변수로 구분하는 것을 특징으로 하는 비염 진단 방법.Rhinitis diagnosis method, characterized in that for calculating a correlation coefficient for each variable of the patient's characteristic information, and dividing a significant variable extracted based on the correlation coefficient as the common variable.
  11. 제 9 항에 있어서,According to claim 9,
    상기 변수 결정 단계는,The variable determination step,
    환자의 위치 정보에 기초하여 획득한 상기 날씨 정보에 포함된 각각의 온도 정보, 습도 정보 및 미세먼지 정보를 상기 개별 변수로 구분하며,Each of the temperature information, humidity information, and fine dust information included in the weather information acquired based on the patient's location information is classified as the individual variable,
    상기 비염 점수에 대한 상기 각각의 온도 정보, 습도 정보 및 미세먼지 정보의 기울기는 환자 별로 상이한 것을 특징으로 하는 비염 진단 방법.Rhinitis diagnosis method, characterized in that the slope of each of the temperature information, humidity information and fine dust information for the rhinitis score is different for each patient.
  12. 제 8 항에 있어서,According to claim 8,
    상기 모델 결정 단계는,The model determination step,
    상기 예측 변수를 기반으로 회귀 모델, 선형 혼합 모델 및 앙상블 모델을 포함하는 복수의 예측 모델을 생성하고, 각 예측 모델의 예측 성능을 판단하여 상기 복수의 예측 모델 중에서 상기 최적화된 예측 모델을 결정하는 것을 특징으로 하는 비염 진단 방법.Generating a plurality of predictive models including a regression model, a linear mixed model, and an ensemble model based on the predictor variables, and determining the optimized predictive model among the plurality of predictive models by determining the predictive performance of each predictive model. Characterized rhinitis diagnosis method.
  13. 제 12 항에 있어서,According to claim 12,
    상기 모델 결정 단계는,The model determination step,
    상기 예측 변수에 해당되는 상기 환자의 특징 정보에 K겹 교차 검증(K fold Cross Validation)을 적용하여 계산한 평균 제곱근 오차(Root Mean Square Error, RMSE)를 기반으로 상기 예측 모델의 예측 성능을 판단하는 것을 특징으로 하는 비염 진단 방법.Based on the Root Mean Square Error (RMSE) calculated by applying K-fold Cross Validation to the patient's feature information corresponding to the predictor variable, Determining the predictive performance of the predictive model Rhinitis diagnosis method, characterized in that.
  14. 제 12 항에 있어서,According to claim 12,
    상기 최적화된 예측 모델은,The optimized predictive model,
    공통 변수 및 개별 변수를 트리 기반의 머신 러닝 알고리즘에 적용하여 생성된 상기 앙상블 모델인 것을 특징으로 하는 비염 진단 방법.Rhinitis diagnosis method, characterized in that the ensemble model generated by applying a common variable and individual variables to a tree-based machine learning algorithm.
  15. 비염 진단 방법을 실행시키기 위한 프로그램을 기록한 기록 매체에 있어서,In the recording medium recording the program for executing the rhinitis diagnosis method,
    비염 중증도 정보, 환경 정보 및 날씨 정보를 포함하는 환자의 특징 정보로부터 상관 분석을 통해 유의한 변수를 추출하여 예측 변수로 결정하는 변수 결정 기능;A variable determination function for extracting significant variables from the patient's characteristic information including rhinitis severity information, environmental information, and weather information through correlation analysis and determining them as predictive variables;
    상기 예측 변수를 기반으로 비염 진단을 위한 예측 모델을 생성하고, 상기 예측 모델의 예측 성능을 판단하여 최적화된 예측 모델을 결정하는 모델 결정 기능; 및 a model determination function for generating a predictive model for diagnosing rhinitis based on the predictive variables and determining an optimized predictive model by determining predictive performance of the predictive model; and
    상기 예측 변수에 해당되는 상기 환자의 특징 정보를 상기 최적화된 예측 모델에 입력하여 환자의 비염 점수를 예측하는 점수 예측 기능을 구현하는 프로그램이 기록되고 컴퓨터로 읽을 수 있는 기록 매체.A computer-readable recording medium on which a program for implementing a score prediction function for predicting the patient's rhinitis score by inputting the patient's characteristic information corresponding to the predictor variable into the optimized predictive model is recorded.
PCT/KR2022/013025 2021-09-07 2022-08-31 Rhinitis diagnosis apparatus, method, and recording medium WO2023038363A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
KR10-2021-0118903 2021-09-07
KR1020210118903A KR20230036261A (en) 2021-09-07 2021-09-07 User equipment, method, and recording medium for diagnosing rhinitis

Publications (1)

Publication Number Publication Date
WO2023038363A1 true WO2023038363A1 (en) 2023-03-16

Family

ID=85502559

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2022/013025 WO2023038363A1 (en) 2021-09-07 2022-08-31 Rhinitis diagnosis apparatus, method, and recording medium

Country Status (2)

Country Link
KR (1) KR20230036261A (en)
WO (1) WO2023038363A1 (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101729694B1 (en) * 2017-01-02 2017-04-25 한국과학기술정보연구원 Method and Apparatus for Predicting Simulation Results
KR102013831B1 (en) * 2017-03-16 2019-08-23 사회복지법인 삼성생명공익재단 Allergic diseases patient-specific environment management system
KR20190115330A (en) * 2018-04-02 2019-10-11 주식회사 씨씨앤아이리서치 An application for predicting an acute exacerbation of chronic respiratory disease
US20200297955A1 (en) * 2015-08-26 2020-09-24 Resmed Sensor Technologies Limited Systems and methods for monitoring and management of chronic disease
WO2021063935A1 (en) * 2019-09-30 2021-04-08 F. Hoffmann-La Roche Ag Prediction of disease status

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200297955A1 (en) * 2015-08-26 2020-09-24 Resmed Sensor Technologies Limited Systems and methods for monitoring and management of chronic disease
KR101729694B1 (en) * 2017-01-02 2017-04-25 한국과학기술정보연구원 Method and Apparatus for Predicting Simulation Results
KR102013831B1 (en) * 2017-03-16 2019-08-23 사회복지법인 삼성생명공익재단 Allergic diseases patient-specific environment management system
KR20190115330A (en) * 2018-04-02 2019-10-11 주식회사 씨씨앤아이리서치 An application for predicting an acute exacerbation of chronic respiratory disease
WO2021063935A1 (en) * 2019-09-30 2021-04-08 F. Hoffmann-La Roche Ag Prediction of disease status

Also Published As

Publication number Publication date
KR20230036261A (en) 2023-03-14

Similar Documents

Publication Publication Date Title
Lombardo et al. ESSENCE II and the framework for evaluating syndromic surveillance systems.
WO2015008936A1 (en) Diagnostic apparatus using habit, diagnosis management apparatus, and diagnostic method using same
WO2020119403A1 (en) Hospitalization data abnormity detection method, apparatus and device, and readable storage medium
WO2019070200A1 (en) A healthcare information management system
US20230409966A1 (en) Training machine learning algorithms with temporaly variant personal data, and applications thereof
Portela et al. Enabling a pervasive approach for intelligent decision support in critical health care
WO2021215809A1 (en) System and method for providing early diagnosis of cognitive disorder and community care matching service for elderly
CN112117002A (en) Novel intelligent psychological assessment intervention system and method combined with virtual reality technology
WO2022211385A1 (en) Health care consultation system using distribution of disease prediction values
US10990900B2 (en) Scheduling machine learning tasks, and applications thereof
Chame et al. SISS-geo: leveraging citizen science to monitor wildlife health risks in Brazil
Pagali et al. Delirium occurrence and association with outcomes in hospitalized COVID-19 patients
WO2023038363A1 (en) Rhinitis diagnosis apparatus, method, and recording medium
McCarron et al. Personal exposure to fine particulate matter (PM2. 5) and self-reported asthma-related health
Wauben et al. Learning from Aviation to Improve Safety in the Operating Room‐a Systematic Literature Review
Beierle et al. Self-assessment of having covid-19 with the corona check mhealth app
Cossin et al. Public Health and Epidemiology Informatics: Recent Research Trends Moving toward Public Health Data Science
CN112711579A (en) Medical data quality detection method and device, storage medium and electronic equipment
Zürcher et al. Novel approach to estimate tuberculosis transmission in primary care clinics in sub-Saharan Africa: protocol of a prospective study
Manuel et al. A commentary on the value of hospital data for covid-19 pandemic surveillance and planning
Martinez Rebollar et al. Automatic detection of social isolation based on human behavior analysis
Ye et al. SCM: a practical tool to implement hospital-based syndromic surveillance
CN112189206A (en) Processing personal data using machine learning algorithms and applications thereof
WO2024123099A1 (en) Personalized symptom management system for atopic dermatitis patients
WO2024128584A1 (en) Artificial intelligence and web-based platform for depression prediction

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: 22867632

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE