CN111985713A - Data index waveform prediction method and device - Google Patents
Data index waveform prediction method and device Download PDFInfo
- Publication number
- CN111985713A CN111985713A CN202010836482.XA CN202010836482A CN111985713A CN 111985713 A CN111985713 A CN 111985713A CN 202010836482 A CN202010836482 A CN 202010836482A CN 111985713 A CN111985713 A CN 111985713A
- Authority
- CN
- China
- Prior art keywords
- index
- predicted
- data
- correlation
- time period
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Business, Economics & Management (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Probability & Statistics with Applications (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a data index waveform prediction method and a device, wherein the method comprises the following steps: determining first index data of an index to be predicted in a time period to be predicted according to historical index data of the index to be predicted; determining index data of each correlation index in a time period to be predicted according to historical index data of each correlation index, wherein the correlation index is a data index having a correlation relation with the index to be predicted; determining second index data of the index to be predicted in the time period to be predicted according to the index data of each associated index in the time period to be predicted; and determining a waveform prediction result of the index to be predicted in the time period to be predicted according to the first index data and the second index data of the index to be predicted in the time period to be predicted. The method and the device combine the historical index data of the index to be predicted and the associated index to predict the index to be predicted, so that the waveform prediction result of the data index is more accurate.
Description
Technical Field
The present invention relates to the field of data processing, and in particular, to a method and an apparatus for predicting a data indicator waveform.
Background
This section is intended to provide a background or context to the embodiments of the invention that are recited in the claims. The description herein is not admitted to be prior art by inclusion in this section.
In the field of operation and maintenance, prediction of waveforms generated by various services and operation and maintenance data is often involved. The existing operation and maintenance data prediction method only analyzes the historical trend of index data to be predicted, and actually, in the operation and maintenance field, the state of certain index data is often internally correlated with the states of other index data. Therefore, how to provide a scheme for predicting the index to be predicted by combining the incidence relation between the index to be predicted and the related index is also an urgent technical problem to be solved at present.
Disclosure of Invention
The embodiment of the invention provides a data index waveform prediction method, which is used for solving the technical problem that the prediction result is inaccurate because the historical waveform trend of a data index to be predicted is only considered in the conventional data index prediction method, and comprises the following steps: determining first index data of an index to be predicted in a time period to be predicted according to historical index data of the index to be predicted; determining index data of each correlation index in a time period to be predicted according to historical index data of each correlation index, wherein the correlation index is a data index having a correlation relation with the index to be predicted; determining second index data of the index to be predicted in the time period to be predicted according to the index data of each associated index in the time period to be predicted; and determining a waveform prediction result of the index to be predicted in the time period to be predicted according to the first index data and the second index data of the index to be predicted in the time period to be predicted.
The embodiment of the invention also provides a data index waveform prediction device, which is used for solving the technical problem that the prediction result is inaccurate because the historical waveform trend of a data index to be predicted is only considered in the conventional data index prediction method, and comprises the following steps: the index prediction module is used for determining first index data of the index to be predicted in the time period to be predicted according to the historical index data of the index to be predicted; the correlation index prediction module is used for determining index data of each correlation index in a time period to be predicted according to historical index data of each correlation index, wherein the correlation index is a data index which has a correlation relation with the index to be predicted; the index data processing module is used for determining second index data of the index to be predicted in the time period to be predicted according to the index data of each associated index in the time period to be predicted; and the waveform prediction module is used for determining a waveform prediction result of the index to be predicted in the time period to be predicted according to the first index data and the second index data of the index to be predicted in the time period to be predicted.
The embodiment of the invention also provides computer equipment for solving the technical problem that the prediction result is inaccurate because the historical waveform trend of the data index to be predicted is only considered in the conventional data index prediction method.
The embodiment of the invention also provides a computer readable storage medium, which is used for solving the technical problem that the prediction result is inaccurate by only considering the historical waveform trend of the data index to be predicted in the conventional data index prediction method.
In the embodiment of the invention, after the first index data of the index to be predicted in the time period to be predicted is determined according to the historical index data of the index to be predicted, the index data of each associated index in the time period to be predicted is determined according to the historical index data of each associated index which has an associated relation with the index to be predicted, the second index data of the index to be predicted in the time period to be predicted is further determined according to the index data of each associated index in the time period to be predicted, finally, the waveform prediction result of the index to be predicted in the time period to be predicted is determined according to the first index data and the second index data of the index to be predicted in the time period to be predicted, compared with the technical scheme that the waveform of the index to be predicted is predicted only considering the historical waveform trend of the index to be predicted in the prior art, the embodiment of the invention combines the index to be predicted and the historical index data of the associated index, and the waveform prediction result of the data index can be more accurate by predicting the index to be predicted.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts. In the drawings:
FIG. 1 is a flow chart illustrating a method for predicting a data indicator waveform according to an embodiment of the present invention;
FIG. 2 is a flow chart of an alternative method for determining a correlation index according to an embodiment of the present invention;
FIG. 3 is a diagram illustrating an apparatus for predicting a data indicator waveform according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of an alternative data indicator waveform prediction apparatus according to an embodiment of the present invention;
fig. 5 is a schematic diagram of a computer device provided in the embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.
An embodiment of the present invention provides a method for predicting a waveform of a data indicator, and fig. 1 is a flowchart of a method for predicting a waveform of a data indicator provided in an embodiment of the present invention, and as shown in fig. 1, the method may include the following steps:
s101, determining first index data of the index to be predicted in the time period to be predicted according to historical index data of the index to be predicted.
It should be noted that the index to be predicted in the embodiment of the present invention may be any data index to be predicted, may be a service data index, and may also be an operation and maintenance data index.
For data in a single operation and maintenance field, the trend is often approximately periodic, different services and different periods, wherein the periods are respectively one day, one week or one month or even one year.
For the periodic data indexes, the index predicted value of the index to be predicted in one or more future periods can be determined according to the index data of the index to be predicted in a plurality of historical periods. Because the data in different history periods have different importance degrees on the data of the prediction result and the data at different history moments have different importance degrees on the prediction result, in general, the farther the data is from the current moment, the smaller the history reference meaning is, and the smaller the influence on the trend and the value of the waveform at the current next moment is.
Therefore, in an embodiment, when the index to be predicted is a periodic data index, the time period to be predicted may include: one or more periods to be predicted of the index to be predicted; the above S101 may be implemented by the following steps: acquiring a weight value corresponding to each history period and a weight value corresponding to each preset moment in each history period; and performing weighted average calculation on the index data of the index to be predicted in each history period according to the weight value corresponding to each history period and the weight value corresponding to each preset moment in each history period to obtain first index data of the index to be predicted in the time period to be predicted. By the embodiment, the historical data in different periods and different moments are endowed with proper weight values to participate in prediction of future data values, and a more accurate prediction result can be obtained.
In specific implementation, the embodiment of the invention introduces a time decay function f1And f2Using f1The weight for representing the influence of the data far away from the time on the data at the current next moment in a period T is f1(t1) Wherein, -T1≤t1Less than or equal to 0; using f2Representing the whole historical data pair in different periodsThe weight of the influence of the data at the previous next moment is f2(t2) Wherein, t2Is a non-positive integer, t20 denotes the current period, t2-1 represents the last cycle). t is t2The | T can be selected according to actual conditions2L values, e.g. considering only the last three cycles of data, then t2Can respectively take the values of 0, -1 and-2.
Calculating the index value DataNextVal of the data index to be predicted at the next moment by the following formula:
wherein the content of the first and second substances,
f1(t)=eα·t(t < ═ 0, α is a constant factor);
f2(t)=1/(t+1)β(t < ═ 0 and is an integer, β is a constant factor);
data (t ═ 0, this function is used to get the wave value at some time in the past);
data is a function of historical Data, Data (0) represents current Data, and Data (t) represents Data from the current time t.
In the embodiment of the invention, periods with different weights are given, different times in each period with different weights, and finally, weighted historical data of each period are superposed to obtain the predictive data PreData of the current next time.
It should be noted that f in the embodiment of the present invention1And f2The formula (2) can be replaced, and the formula with the best effect can be selected according to different specific application situations.
And S102, determining index data of each correlation index in a time period to be predicted according to the historical index data of each correlation index, wherein the correlation index is a data index which has a correlation relation with the index to be predicted.
It should be noted that, in an actual application scenario, a change of one data index often has a certain correlation with other data indexes, and therefore, by analyzing index data of some correlation indexes, prediction of an index to be predicted can be facilitated.
It should be noted that the data index having an association relation with the index to be predicted may be a periodic data index, a non-periodic data index, or a discrete data index, and then the above S102 may be implemented by any one or more of the following ways:
the first method is as follows: if the associated index is a periodic data index, determining index data of the associated index in a time period to be predicted according to the index data of the associated index in a plurality of historical periods;
the second method comprises the following steps: if the correlation index is an aperiodic data index, determining index data of the correlation index in a time period to be predicted based on the long-term and short-term memory network model;
the third method comprises the following steps: and if the associated index is a discrete data index, determining index data of the associated index in the time period to be predicted based on a naive Bayes classification model.
S103, according to the index data of each associated index in the time period to be predicted, determining second index data of the index to be predicted in the time period to be predicted.
In specific implementation, second index data of the index to be predicted in the time period to be predicted can be determined according to the index data of each associated index in the time period to be predicted based on a pre-trained AdaBoost algorithm regression model.
And S104, determining a waveform prediction result of the index to be predicted in the time period to be predicted according to the first index data and the second index data of the index to be predicted in the time period to be predicted.
As can be seen from the above, in the waveform prediction method of the data index provided in the embodiment of the present invention, after the first index data of the to-be-predicted index in the to-be-predicted time period is determined according to the historical index data of each associated index having an association relationship with the to-be-predicted index, the index data of each associated index in the to-be-predicted time period is determined according to the historical index data of each associated index having an association relationship with the to-be-predicted index, the second index data of the to-be-predicted index in the to-be-predicted time period is further determined according to the index data of each associated index in the to-be-predicted time period, and finally, the waveform prediction result of the to-be-predicted index in the to-be-predicted time period is determined according to the first index.
By the waveform prediction method of the data index provided by the embodiment of the invention, the index to be predicted is predicted by combining the historical index data of the index to be predicted and the associated index, so that the waveform prediction result of the data index is more accurate.
In an embodiment, as shown in fig. 2, the waveform prediction method for a data index provided in the embodiment of the present invention may determine each associated index having an associated relationship with an index to be predicted by the following steps:
s201, acquiring a pre-configured association index mapping table, wherein the association index mapping table comprises association relations among data indexes;
s202, determining a correlation index set of the indexes to be predicted according to the correlation index mapping table, wherein the correlation index set comprises: and one or more associated indexes having an associated relation with the index to be predicted.
Table 1 shows a related index mapping table, assuming that X represents a waveform of an index to be measured, Y represents a set of operation and maintenance indexes related to the index to be measured, and Yi represents the ith index in the set; and respectively carrying out data prediction on each index in the Y. If the relevant index set has periodic waveform, non-periodic waveform and discrete data index, predicting the periodic waveform by the same method as the first part; if the waveform is a non-periodic waveform, predicting through LTSM; if the index data is discrete index data, the prediction is carried out through a multi-classification model trained through naive Bayes. Finally, the predicted value Ypre of Y is obtained.
When determining the index data (i.e., the second index data) of the index to be predicted in the time period to be predicted according to the index data of each associated index in the time period to be predicted, it may be assumed that the value of X at time t is Xt and the value of Y is Yt (Yt is a set). For historical data of X and Y, taking Xt and Yt corresponding to a plurality of time points to form a training set; using an AdaBoost regression model, taking Yt as model input and Xt as model output, and training the model to obtain a model M-ada based on Y prediction X; after the model M-ada is trained, inputting the predicted value Ypre of the associated index into the model M-ada to obtain an index value Xpre of the index to be predicted; and finally, averaging the PreData and the Xpre to obtain a predicted value of the index of the data to be predicted, namely X-fin ═ PreData + Xpre)/2.
In specific implementation, the data index waveform prediction method provided in the embodiment of the present invention may be implemented by the following steps:
the first step is as follows: calling an N6 module, calling an N6 module and calling an N0 module, wherein the relationship among the modules in the N0 is as follows:
calling M1 to determine whether the index to be measured is a periodic data index, and if so, determining the period T1.
And calling M2 to determine a time attenuation function f1 to be used and determine constant factors in the time attenuation function.
Calling M3 to determine the time attenuation function f2 to be used and determine the constant factor.
Invoking M4, invoking M5 by M4 (invoking M7 at the bottom layer of M5) to read historical data, and simultaneously invoking M2 to obtain f1Call M3 to get f2Assist M4 in performing online prediction (the formula used for online prediction is the formula for solving DataNextVal described above). And outputting the predicted value of the data wave at the next moment. If the data wave value of a future period of time is not needed to be predicted, the operation is exited, otherwise, the operation is entered.
Determining the time length F of a section of data wave needing to be predicted in the future. And (4) calling M4 iteratively (simulating historical data by using M4 predicted data in each round), so that predicted data waves of a future period F are obtained iteratively, and a predicted result is output.
The second step is that: and calling an N3 module to determine a correlation index set of the indexes to be measured.
The third step: calling N4, and calling N0, N1 and N2 by the N4 according to the association index set to respectively obtain the predicted values of the association indexes.
The fourth step: the AdaBoost model of N5 was called (the model needs to be trained well in advance based on AdaBoost).
The fifth step: calling N7, receiving the prediction results output by N5 and N6, and outputting the final prediction result after integration.
Wherein M1 is a period determination module; m2 is f1A time decay function module; m3 is f2A time decay function module; m4 is an online prediction function module; m5 is a historical data reading module; m6 is a future segment data wave value prediction module; m7 is a database module; n0 is a periodic continuous data prediction module; n1 is a non-periodic continuous data prediction module (LSTM); n2 is a discrete data prediction module (naive bayes); n3 is a reading module of the associated index mapping table; n4 is a correlation index prediction module; n5 is a module for predicting an index to be predicted based on the correlation index; n6 is a module for predicting the index to be predicted based on the historical data of the index to be predicted; n7 is a result integration module.
The functions of the respective modules are explained below:
(1) module N0:
m1 receives historical data and service knowledge and outputs a data wave period T1. To M4. Outputs f1 and f2To M2 and M3, respectively;
m2 receives t1Characterizing the position of a history data in a cycle, and outputting the weight f of the history data corresponding to the position1(t1). To M4;
m3 receives t2Representing a historical data, separated from the current data by T2 cycles, and outputting the weight f of the whole historical data belonging to the cycle2(t2) To M4;
m5 receives the time t of a certain past moment, calls M7, outputs the historical Data of the Data wave at the moment and transmits the historical Data to M4;
m4 inputs the data wave period T1 and the history period number T2 which needs to be traced back, calls M5, M2 and M3 and outputs the predicted value of the data wave at the next moment. If M4 is called by M6, the result is transmitted to M6, otherwise, the terminal is directly output;
inputting a parameter F needing to predict how long a period of time will come by the M6, and calling the M4 in an iterative manner to obtain a data wave continuous value with the length of F in a period of time;
the input of M7 is t, which represents the time length from the past time to the current time, outputs the data wave value from the current time to the history, and transmits the result to M5.
Modules N1, N2: in the initialization phase: receiving the associated index mapping table of N3, and respectively training a prediction model for all related indexes; in the service phase: and receiving each correlation index of the indexes to be measured, calling a correlation model, and outputting a prediction result of each correlation index.
Module N3: acquiring a correlation index mapping table configured in advance by operation and maintenance personnel;
module N4: and receiving each relevant index of the indexes to be measured, and calling the corresponding N1 and N2 models for prediction. And taking the return values of N1 and N2 as output results.
Module N5: in the initialization phase: receiving an associated index mapping table of N3, and training corresponding AdaBoost models for each row in the table according to the relation between two rows of data in the table; in the service phase: and receiving the output of the N4, calling a relevant model, and outputting a predicted value of the index to be measured.
Module N6: calling N0 to obtain a predicted value.
Module N7: and receiving the outputs of N5 and N6, and outputting the final result after integration.
Based on the same inventive concept, the embodiment of the present invention further provides a data indicator waveform predicting apparatus, as described in the following embodiments. Because the principle of the device for solving the problems is similar to the data index waveform prediction method, the implementation of the device can refer to the implementation of the data index waveform prediction method, and repeated details are not repeated.
Fig. 3 is a schematic diagram of an apparatus for predicting a data indicator waveform according to an embodiment of the present invention, as shown in fig. 3, the apparatus may include: an index prediction module 31, a correlation index prediction module 32, an index data processing module 33, and a waveform prediction module 34.
The index prediction module 31 is configured to determine, according to historical index data of an index to be predicted, first index data of the index to be predicted in a time period to be predicted; the correlation index prediction module 32 is configured to determine index data of each correlation index in a to-be-predicted time period according to historical index data of each correlation index, where the correlation index is a data index having a correlation relationship with the to-be-predicted index; the index data processing module 33 is configured to determine, according to the index data of each associated index in the to-be-predicted time period, second index data of the to-be-predicted index in the to-be-predicted time period; and the waveform predicting module 34 is configured to determine a waveform prediction result of the to-be-predicted index in the to-be-predicted time period according to the first index data and the second index data of the to-be-predicted index in the to-be-predicted time period.
As can be seen from the above, in the waveform prediction apparatus for data indexes provided in the embodiment of the present invention, the index prediction module 31 determines, according to the historical index data of the index to be predicted, first index data of the index to be predicted in the time period to be predicted; determining index data of each correlation index in the time period to be predicted according to historical index data of each correlation index having a correlation relation with the index to be predicted by a correlation index prediction module 32; determining second index data of the index to be predicted in the time period to be predicted according to the index data of each associated index in the time period to be predicted by the index data processing module 33; and determining a waveform prediction result of the index to be predicted in the time period to be predicted according to the first index data and the second index data of the index to be predicted in the time period to be predicted by the waveform prediction module 34.
By the waveform prediction device of the data index provided by the embodiment of the invention, the index to be predicted is predicted by combining the historical index data of the index to be predicted and the associated index, so that the waveform prediction result of the data index can be more accurate.
In an embodiment, in the data index waveform predicting apparatus provided in the embodiment of the present invention, the index to be predicted may be a periodic data index, and the time period to be predicted may include: one or more periods to be predicted of the index to be predicted; as shown in fig. 4, the index prediction module 31 may include: the weight configuration module 311 is configured to obtain a weight value corresponding to each history period and a weight value corresponding to each preset time in each history period; the index calculating module 312 is configured to perform weighted average calculation on the index data of the to-be-predicted index in each history cycle according to the weight value corresponding to each history cycle and the weight value corresponding to each preset time in each history cycle, so as to obtain first index data of the to-be-predicted index in the to-be-predicted time period.
In one embodiment, as shown in fig. 4, the correlation index prediction module 32 includes: a first associated index prediction module 321, configured to determine, according to index data of the associated index in multiple historical periods, index data of the associated index in a time period to be predicted, if the associated index is a periodic data index; a second correlation index prediction module 322, configured to determine, based on the long-short term memory network model, index data of the correlation index in the time period to be predicted if the correlation index is an aperiodic data index; and the third correlation index prediction module 323 is used for determining index data of the correlation index in the time period to be predicted based on the naive Bayesian classification model.
In an embodiment, as shown in fig. 4, the data index waveform predicting apparatus provided in the embodiment of the present invention may further include: a correlation index mapping table obtaining module 35, configured to obtain a pre-configured correlation index mapping table, where the correlation index mapping table includes a correlation relationship between each data index; a correlation index determining module 36, configured to determine a correlation index set of the to-be-predicted index according to the correlation index mapping table, where the correlation index set includes: and one or more associated indexes having an associated relation with the index to be predicted.
Optionally, the index data processing module 33 may be further configured to determine, based on a pre-trained regression model of the AdaBoost algorithm, second index data of the to-be-predicted index in the to-be-predicted time period according to the index data of each associated index in the to-be-predicted time period.
Based on the same inventive concept, an embodiment of the present invention further provides a computer device, so as to solve the technical problem that the prediction result is inaccurate due to only considering the historical waveform trend of the data index to be predicted in the conventional data index prediction method, and as shown in fig. 5, fig. 5 is a schematic diagram of a computer device provided in an embodiment of the present invention, as shown in fig. 5, the computer device 50 includes a memory 501, a processor 502, and a computer program stored in the memory 501 and operable on the processor 502, and the processor 502 implements the data index waveform prediction method when executing the computer program.
Based on the same inventive concept, the embodiment of the present invention further provides a computer-readable storage medium, so as to solve the technical problem that the prediction result is inaccurate due to the fact that only the historical waveform trend of the data index to be predicted is considered in the existing data index prediction method, and the computer-readable storage medium stores a computer program for executing the data index waveform prediction method.
In summary, embodiments of the present invention provide a method and an apparatus for predicting a waveform of a data index, a computer device, and a computer-readable storage medium, and compared with the technical solution in the prior art that only a historical waveform trend of a data index to be predicted is considered to predict a waveform of the data index to be predicted, the embodiments of the present invention combine historical index data of the index to be predicted and associated indexes to predict the index to be predicted, so that a waveform prediction result of the data index can be more accurate.
It should be noted that, in the prior art, the waveform trend of the data index is predicted by using the neural network model, for the data wave with periodic trend, the interpretability of the neural network model is poor, no clear logic exists, and once the model is trained, the model is used for prediction, so that the flexibility is lacked; if the reference range of the historical data needs to be expanded, the model needs to be retrained, but the training of deep learning is relatively slow, so that the response capability of quickly changing the configuration is lacked, in addition, the operation of the neural network model is complex, the requirement on hardware is too high, and the terminal equipment cannot be deployed necessarily. The data index waveform prediction scheme provided by the embodiment of the invention determines the periodThe method depends on the real observation and analysis of the objective world, makes full use of the prior knowledge of the business logic for confirmation, and has strong interpretability; if the reference range of the historical data needs to be expanded, only the time attenuation function f needs to be modified2The value range of the independent variable can be quickly realized, and the hardware requirement is low because only mathematical operation is involved; the information of the associated indexes is used for assisting the prediction of the indexes to be measured, so that the prediction result is more accurate.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (12)
1. A method for predicting a data indicator waveform, comprising:
determining first index data of an index to be predicted in a time period to be predicted according to historical index data of the index to be predicted;
determining index data of each correlation index in the time period to be predicted according to historical index data of each correlation index, wherein the correlation index is a data index having a correlation relation with the index to be predicted;
according to the index data of each associated index in the time period to be predicted, determining second index data of the index to be predicted in the time period to be predicted;
and determining a waveform prediction result of the index to be predicted in the time period to be predicted according to the first index data and the second index data of the index to be predicted in the time period to be predicted.
2. The method of claim 1, wherein the index to be predicted is a periodic data index, and the time period to be predicted comprises: one or more periods to be predicted of the index to be predicted; the method for determining the first index data of the index to be predicted in the time period to be predicted according to the historical index data of the index to be predicted comprises the following steps:
acquiring a weight value corresponding to each history period and a weight value corresponding to each preset moment in each history period;
and performing weighted average calculation on the index data of the index to be predicted in each history period according to the weight value corresponding to each history period and the weight value corresponding to each preset moment in each history period to obtain first index data of the index to be predicted in the time period to be predicted.
3. The method of claim 1, wherein determining the index data of each relevant index in the time period to be predicted according to the historical index data of each relevant index comprises:
if the associated index is a periodic data index, determining index data of the associated index in the time period to be predicted according to the index data of the associated index in a plurality of historical periods;
if the correlation index is an aperiodic data index, determining index data of the correlation index in the time period to be predicted based on a long-term and short-term memory network model;
and if the associated index is a discrete data index, determining index data of the associated index in the time period to be predicted based on a naive Bayes classification model.
4. The method of claim 3, wherein the method further comprises:
acquiring a pre-configured association index mapping table, wherein the association index mapping table comprises association relations among all data indexes;
determining a correlation index set of indexes to be predicted according to the correlation index mapping table, wherein the correlation index set comprises: and one or more associated indexes having an associated relation with the index to be predicted.
5. The method of claim 1, wherein determining second index data of the index to be predicted in the time period to be predicted according to the index data of each associated index in the time period to be predicted comprises:
and determining second index data of the index to be predicted in the time period to be predicted according to the index data of each associated index in the time period to be predicted based on a pre-trained AdaBoost algorithm regression model.
6. A data indicator waveform prediction apparatus, comprising:
the index prediction module is used for determining first index data of the index to be predicted in the time period to be predicted according to historical index data of the index to be predicted;
the correlation index prediction module is used for determining index data of each correlation index in the time period to be predicted according to historical index data of each correlation index, wherein the correlation index is a data index which has a correlation relation with the index to be predicted;
the index data processing module is used for determining second index data of the index to be predicted in the time period to be predicted according to the index data of each associated index in the time period to be predicted;
and the waveform prediction module is used for determining a waveform prediction result of the index to be predicted in the time period to be predicted according to the first index data and the second index data of the index to be predicted in the time period to be predicted.
7. The apparatus of claim 6, wherein the metric to be predicted is a periodic data metric, and the time period to be predicted comprises: one or more periods to be predicted of the index to be predicted; the index prediction module includes:
the weight configuration module is used for acquiring a weight value corresponding to each history period and a weight value corresponding to each preset moment in each history period;
and the index calculation module is used for performing weighted average calculation on the index data of the index to be predicted in each history period according to the weight value corresponding to each history period and the weight value corresponding to each preset moment in each history period to obtain first index data of the index to be predicted in the time period to be predicted.
8. The apparatus of claim 6, wherein the correlation metric prediction module comprises:
the first correlation index prediction module is used for determining index data of the correlation index in the time period to be predicted according to the index data of the correlation index in a plurality of historical periods if the correlation index is a periodic data index;
the second correlation index prediction module is used for determining index data of the correlation index in the time period to be predicted based on the long-short term memory network model if the correlation index is an aperiodic data index;
and the third correlation index prediction module is used for determining index data of the correlation index in the time period to be predicted based on a naive Bayesian classification model.
9. The apparatus of claim 6, wherein the apparatus further comprises:
the system comprises a correlation index mapping table acquisition module, a correlation index mapping table acquisition module and a correlation index mapping table acquisition module, wherein the correlation index mapping table acquisition module is used for acquiring a pre-configured correlation index mapping table, and the correlation index mapping table comprises the correlation relation among all data indexes;
a correlation index determining module, configured to determine a correlation index set of an index to be predicted according to the correlation index mapping table, where the correlation index set includes: and one or more associated indexes having an associated relation with the index to be predicted.
10. The device of claim 6, wherein the index data processing module is further configured to determine, based on a pre-trained AdaBoost algorithm regression model, second index data of the index to be predicted in the time period to be predicted according to index data of each associated index in the time period to be predicted.
11. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the data index waveform prediction method of any one of claims 1 to 5 when executing the computer program.
12. A computer-readable storage medium storing a computer program for executing the data index waveform prediction method according to any one of claims 1 to 5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010836482.XA CN111985713B (en) | 2020-08-19 | 2020-08-19 | Data index waveform prediction method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010836482.XA CN111985713B (en) | 2020-08-19 | 2020-08-19 | Data index waveform prediction method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111985713A true CN111985713A (en) | 2020-11-24 |
CN111985713B CN111985713B (en) | 2023-08-18 |
Family
ID=73435424
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010836482.XA Active CN111985713B (en) | 2020-08-19 | 2020-08-19 | Data index waveform prediction method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111985713B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103617564A (en) * | 2013-12-04 | 2014-03-05 | 国家电网公司 | Whole-network load prediction method based on local load predicted value comprehensive evaluation |
CN106156875A (en) * | 2015-04-10 | 2016-11-23 | 株式会社日立制作所 | For the method and apparatus that prediction object is predicted |
CN106549772A (en) * | 2015-09-16 | 2017-03-29 | 华为技术有限公司 | Resource prediction method, system and capacity management device |
WO2019056499A1 (en) * | 2017-09-20 | 2019-03-28 | 平安科技(深圳)有限公司 | Prediction model training method, data monitoring method, apparatuses, device and medium |
CN109872003A (en) * | 2019-03-06 | 2019-06-11 | 中国科学院软件研究所 | Obj State prediction technique, system, computer equipment and storage medium |
-
2020
- 2020-08-19 CN CN202010836482.XA patent/CN111985713B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103617564A (en) * | 2013-12-04 | 2014-03-05 | 国家电网公司 | Whole-network load prediction method based on local load predicted value comprehensive evaluation |
CN106156875A (en) * | 2015-04-10 | 2016-11-23 | 株式会社日立制作所 | For the method and apparatus that prediction object is predicted |
CN106549772A (en) * | 2015-09-16 | 2017-03-29 | 华为技术有限公司 | Resource prediction method, system and capacity management device |
WO2019056499A1 (en) * | 2017-09-20 | 2019-03-28 | 平安科技(深圳)有限公司 | Prediction model training method, data monitoring method, apparatuses, device and medium |
CN109872003A (en) * | 2019-03-06 | 2019-06-11 | 中国科学院软件研究所 | Obj State prediction technique, system, computer equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN111985713B (en) | 2023-08-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110686633B (en) | Landslide displacement prediction method and device and electronic equipment | |
KR101919076B1 (en) | Time-series data predicting system | |
CN109472412A (en) | A kind of prediction technique and device of event | |
CN109272142A (en) | Data predication method, device, computer equipment and storage medium based on event | |
CN111461445B (en) | Short-term wind speed prediction method and device, computer equipment and storage medium | |
EP3239905A1 (en) | Methods and apparatus for use in predicting non-stationary time-series data | |
CN115249081A (en) | Object type prediction method and device, computer equipment and storage medium | |
CN114417733A (en) | Method and device for constructing power consumption prediction model, electronic equipment and storage medium | |
Yanto et al. | Implementation multiple linear regresion in neural network predict gold price | |
CN111563236A (en) | Short-term wind speed prediction method and device | |
Salvi et al. | Long short-term model for Brent oil price forecasting | |
CN116542328B (en) | Knowledge distillation method and device for CTR prediction model | |
CN111985713B (en) | Data index waveform prediction method and device | |
Mohammed et al. | Deep learning and neural network-based wind speed prediction model | |
CN110428086A (en) | Prediction technique and device | |
CN114612231A (en) | Stock quantitative trading method and device, terminal device and readable storage medium | |
Bakumenko et al. | Synthesis method of robust neural network models of systems and processes | |
Kim et al. | Data-driven prognostics | |
CN113159419A (en) | Group feature portrait analysis method, device and equipment and readable storage medium | |
Crone et al. | Feature selection of autoregressive neural network inputs for trend time series forecasting | |
He et al. | Singular-Value Decomposition Feature-Extraction Method for Cost-Performance Prediction | |
US20230267363A1 (en) | Machine learning with periodic data | |
CN113379455B (en) | Order quantity prediction method and equipment | |
CN112990826B (en) | Short-time logistics demand prediction method, device, equipment and readable storage medium | |
CN114492212B (en) | Dynamic capacity increasing method and device for low-voltage distribution network, terminal and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |