WO2021151304A1 - Method and apparatus for hysteretic processing of time series data, electronic device, and storage medium - Google Patents

Abstract

A method and an apparatus for hysteretic processing of time series data, an electronic apparatus, and a storage medium relating to data processing technology. The method comprises: performing time series differencing processing and curve fitting on time sequence sample data to obtain a real curve; using a pre-constructed plurality of time series analysis models to perform time series differencing processing and curve fitting on the time series sample data to generate a plurality of predicted curves; on the basis of loss values between the predicted curves and the real curve, calculating degrees of hysteresis of the predicted curves, on the basis of the degrees of hysteresis, filtering a standard time series analysis model out from the plurality of time series analysis models, and using the standard time series analysis model to perform data analysis on time series data of a pre-set event to obtain a trend of development of the pre-set event. The present method is able to improve the accuracy of prediction in time series analysis processes.

Description

Time series data hysteresis processing method, device, electronic equipment and storage medium

This application claims the priority of the Chinese patent application filed with the Chinese Patent Office on July 9, 2020, the application number is 202010656657.9, and the invention title is "Time-sequential data lag processing method, device, electronic equipment and storage medium", and its entire content Incorporated in this application by reference.

Technical field

This application relates to the field of data processing technology, and in particular to a method, device, electronic equipment, and computer-readable storage medium for processing time series data with hysteresis.

Background technique

Lag refers to the lag between one phenomenon and another closely related phenomenon. This situation is usually caused by the influence of itself or other variables. The lag is used to measure the degree of lag, the degree of lag Evaluation refers to the use of a time series analysis model to solve the problem of lag. For example, there will be a lag in the method of using information related to the influenza epidemic to predict the intensity of the influenza next week. Due to the high variability and uncertainty of influenza viruses, influenza prevention and control faces many challenges. The lagging assessment of influenza epidemics has become a key strategy for influenza prevention and control.

The inventor realizes that when the acquired information samples are getting larger and larger, the traditional time series analysis model predicts the next period of information based on the historical information extracted from the information samples, there will be a long-term lag phenomenon, which will cause the calculation process The complexity of the assessment and the decline in the accuracy of the assessment.

Summary of the invention

A method for processing time series data hysteresis provided by this application includes:

Acquiring time series sample data, and performing time series differential processing on the time series sample data to obtain a differential data set;

Curve fitting the difference data set to obtain a true curve;

Through pre-built multiple time series analysis models, analyze the time series sample data to obtain multiple time series analysis data, perform time series differential processing on the multiple time series analysis data, obtain multiple differential analysis data sets, and analyze multiple differential analysis data. Perform curve fitting on the data set to obtain multiple prediction curves;

Using a loss function to calculate a plurality of loss values between the predicted curve and the real curve;

Calculating the lag factor of each prediction curve based on the loss value, and determining the lag of each prediction curve according to the lag factor;

Filtering out a standard time series analysis model from the plurality of time series analysis models according to the lag of each of the prediction curves;

The standard time series analysis model is used to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.

The present application also provides a method and apparatus for processing time series data lag, the apparatus includes:

A differential data generation module, used to obtain time series sample data, and perform time series differential processing on the time series sample data to obtain a differential data set;

The true curve generation module is used to perform curve fitting on the difference data set to obtain a true curve;

The prediction curve generation module is used to analyze the time series sample data through multiple pre-built time series analysis models to obtain multiple time series analysis data, and perform time series differential processing on the multiple time series analysis data to obtain multiple differential analysis data Set, curve fitting multiple differential analysis data sets to obtain multiple prediction curves;

A loss value calculation module, configured to use a loss function to calculate a plurality of loss values between the predicted curve and the true curve;

A hysteresis calculation module, configured to calculate the hysteresis factor of each prediction curve based on the loss value, and determine the hysteresis of each prediction curve according to the hysteresis factor;

The standard time series analysis model screening module, according to the hysteresis of each of the prediction curves, selects standard time series analysis models from the plurality of time series analysis models;

The data analysis module is configured to use the standard time series analysis model to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.

This application also provides an electronic device, which includes:

Memory, storing at least one instruction; and

The processor executes the instructions stored in the memory to implement the following steps:

Acquiring time series sample data, and performing time series differential processing on the time series sample data to obtain a differential data set;

Curve fitting the difference data set to obtain a true curve;

Through pre-built multiple time series analysis models, analyze the time series sample data to obtain multiple time series analysis data, perform time series differential processing on the multiple time series analysis data, obtain multiple differential analysis data sets, and analyze multiple differential analysis data. Perform curve fitting on the data set to obtain multiple prediction curves;

Using a loss function to calculate a plurality of loss values between the predicted curve and the real curve;

Calculating the lag factor of each prediction curve based on the loss value, and determining the lag of each prediction curve according to the lag factor;

Filtering out a standard time series analysis model from the plurality of time series analysis models according to the lag of each of the prediction curves;

The standard time series analysis model is used to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.

This application also provides a computer-readable storage medium that stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

Acquiring time series sample data, and performing time series differential processing on the time series sample data to obtain a differential data set;

Curve fitting the difference data set to obtain a true curve;

Through pre-built multiple time series analysis models, analyze the time series sample data to obtain multiple time series analysis data, perform time series differential processing on the multiple time series analysis data, obtain multiple differential analysis data sets, and analyze multiple differential analysis data. Perform curve fitting on the data set to obtain multiple prediction curves;

Using a loss function to calculate a plurality of loss values between the predicted curve and the real curve;

Calculating the lag factor of each prediction curve based on the loss value, and determining the lag of each prediction curve according to the lag factor;

Filtering out a standard time series analysis model from the plurality of time series analysis models according to the lag of each of the prediction curves;

The standard time series analysis model is used to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.

Description of the drawings

FIG. 1 is a schematic flowchart of a method for processing time-series data hysteresis according to an embodiment of the application;

2 is a schematic flowchart of one of the steps in the method for processing time series data hysteresis according to an embodiment of the application;

3 is a schematic diagram of modules of a time series data hysteresis processing device provided by an embodiment of the application;

4 is a schematic diagram of the internal structure of an electronic device for implementing a method for processing time series data hysteresis according to an embodiment of the application;

The realization, functional characteristics, and advantages of the purpose of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings.

Detailed ways

It should be understood that the specific embodiments described here are only used to explain the present application, and are not used to limit the present application.

The execution subject of the time-series data hysteresis processing method provided by the embodiment of the present application includes but is not limited to at least one of the electronic devices that can be configured to execute the method provided by the embodiment of the present application, such as a server and a terminal. In other words, the time-series data hysteresis processing method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

Referring to FIG. 1, it is a schematic flowchart of a method for processing time series data hysteresis according to an embodiment of this application.

In this embodiment, the method for processing time series data hysteresis includes:

S1. Acquire time series sample data, and perform time series differential processing on the time series sample data to obtain a differential data set.

In the embodiments of the present application, the time-series sample data is also called time-series sample data, and includes any data with time-series characteristics, such as data of epidemic intensity changes over time, data of object temperature changes over time, and so on. In the embodiment of the present application, a java statement with a data calling function may be used to obtain the time series sample data from a database for storing the time series sample data.

Preferably, before the time-series differential processing is performed on the time-series sample data, the method further includes:

Judging whether the time series sample data has stationarity;

When the time series sample data has stationarity, do not perform time series difference processing on the time series sample data;

When the time series sample data does not have stationarity, the time series difference processing of the time series sample data is performed.

Preferably, the embodiment of the present application can determine whether the time series sample data has stationarity by detecting whether there is a unit root in the time series sample data. Wherein, the unit root refers to the root modulo 1 in the time series sample data, for example: 1, -1, i, and -i are all 4th unit roots.

When there is no unit root in the time series sample data, it means that the time series sample data has stationarity, and when there is a unit root in the time series sample data, it means that the time series sample data does not have stationarity.

Preferably, the embodiment of the present application adopts a first-order difference to perform a time-series difference on the time-series sample data, wherein the first-order difference formula is as follows:

Δ _yx =y _x+1 -y _x

Wherein, _Δyx is the first-order difference of the object temperature change data with time, y _x+1 is the data at the time x+1, and y _x is the data at the time x.

In the embodiment of the present application, performing time-series difference processing on the time-series sample data can filter out unstable factors in the time-series sample data.

S2. Perform curve fitting on the differential data set to obtain a true curve.

Specifically, referring to FIG. 2, the detailed implementation process of S2 includes:

S20. Extract the time sequence characteristics of the differential data in the differential data set;

S21: Use the time sequence feature to perform coordinate encoding on the differential data in the differential data set;

S22. Use the coordinate encoding to map the difference data in the difference data set to a pre-built coordinate system;

S23. Connect the differential data in the coordinate system according to the time series feature to obtain the true curve. The true curve is a stable curve, which can represent the curve of the difference data with time.

In the embodiment of the present application, the time sequence feature refers to the time sequence of the differential data; the coordinate encoding of the differential data in the differential data set according to the time sequence feature includes the chronological sequence of the differential data. The difference data is expressed in a preset coordinate system.

For example, establish a coordinate system with time t and t+1 as the horizontal axis, and L _t and L _t+1 as the vertical axis to establish a coordinate system; In the coordinate system, the difference data in the coordinate system is further connected according to the time sequence feature to obtain the true curve.

S3. Analyze the time series sample data through multiple pre-built time series analysis models to obtain multiple time series analysis data, and perform time series differential processing on the multiple time series analysis data to obtain multiple differential analysis data sets. Curve fitting is performed on the differential analysis data set to obtain multiple prediction curves.

In the embodiment of the present application, the time sequence analysis model is:

S _t+1 =X _t +Y _t

X _t =αS _t +(1-α)(S _t-1 -Y _t-1 )

Y _t ＝β(X _t -X _t-1 )+(1-β)Y _t-1

Among them, _St represents the difference data of the difference data concentrated at time t, _St-1 represents the difference data of the difference data concentrated at time t-1, X _t represents the smooth value at time t, and X _t-1 represents The smoothed value at time t-1, Y _t represents the trend value at time t, Y _t-1 represents the trend value at time t-1, and S _t+1 represents that the differential analysis data is concentrated in the differential analysis data at time t+1 , Α and β are preset different smoothing coefficients.

The embodiment of the present application obtains multiple time series analysis models by customizing different smoothing coefficients α and β.

Specifically, the embodiment of the present application uses the multiple time series analysis models to analyze the time series sample data to obtain multiple time series analysis data. For example, for the data that the object temperature changes with time as the time series sample data: [(1,2), (2,4), (3,5), (4,8)], where the first digit is the time The identifier is used to indicate the time of the time series sample data, the second digit is the data identifier, which is used to indicate the numerical value of the object temperature that changes with time; The changed data is analyzed, and multiple time series analysis data are obtained: [(5,9), (6,11), (7,14), (8,16)].

Further, since the time series analysis data may contain unstable factors, the embodiment of the present application further performs time series differential processing on the time series analysis data to filter out the unstable factors in the time series analysis data. .

In detail, the embodiment of the present application performs the time series differential processing as described in S1 on the multiple time series analysis data to obtain multiple differential analysis data sets.

In the embodiment of the present application, curve fitting is performed on multiple differential analysis data sets to obtain multiple prediction curves.

Preferably, the method of performing curve fitting on the multiple differential analysis data sets is the same as the method of performing curve fitting on the differential data set in S2, and will not be repeated here.

S4. Calculate the loss value between the multiple predicted curves and the real curve by using a loss function.

Preferably, the embodiment of the present application uses the following formula to calculate the initial loss value between the multiple predicted curves and the true curve:

in,

Represents the initial loss value,

It represents the predicted value of the prediction curve, σ represents the number of the time series analysis models, Y represents the true value of the true curve under different translation degrees, and α represents the error factor.

Further, calculating the root mean square error of the initial loss value according to the initial loss value to obtain the loss value includes:

in,

Is the root mean square error of the initial loss value, that is, the loss value,

Represents the initial loss value, and m represents the number of time series analysis models.

S5. Calculate the lag factors of multiple prediction curves based on the loss value, and determine the lag degrees of the multiple prediction curves according to the lag factors.

In the embodiment of the present application, the lag factor τ _m is calculated using the following algorithm:

Among them, arg min refers to the set of all independent variables that make the function obtain its minimum value,

It is an analytical value, τ is a loss value, i is the starting value of the prediction curve, and n is the end value of the prediction curve.

Further, the embodiment of the present application determines the hysteresis of each prediction curve according to the hysteresis factor, including:

Use the following algorithm to calculate the lag score _delay :

Wherein, τ _m is the lag factor of the prediction curve;

It is the smallest root mean square error after _{τ m translation.}

In the embodiment of the present application, the lag factors of multiple prediction curves are substituted into the above algorithm to obtain the lag of multiple prediction curves. The lag indicates the degree of lag in comparison between the prediction curve and the real curve.

S6. According to the hysteresis of each of the prediction curves, a standard time series analysis model is selected from the plurality of time series analysis models.

In the embodiment of the present application, the selection of a standard time series analysis model from the multiple time series analysis models according to the hysteresis of each of the prediction curves includes:

The multiple time series analysis models are screened according to the hysteresis, and a standard time series analysis model is obtained.

In the embodiment of the present application, models with hysteresis less than a preset threshold among multiple time series analysis models are screened out to obtain a standard time series analysis model.

The standard time series analysis model is a model selected after comparing the lag and error degree between different time series analysis models, which can solve the long-term lag problem in time series forecasting. The lag is used to measure the degree of lag If the lag of the multiple time series analysis models is smaller, the data processed by the standard time series analysis model will be relatively accurate, and there will be fewer lag problems.

In life, influenza viruses are highly variable and uncertain. Correctly predicting the peak of influenza epidemics is critical to preventing and controlling influenza. The selected standard time series analysis model can solve the problems of lagging forecast results and large forecast errors when forecasting influenza peaks, and improve the accuracy of influenza forecasting.

S7. Use the standard time series analysis model to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.

In the embodiment of the present application, the time sequence to be processed is analyzed according to the standard time sequence analysis model obtained through screening to obtain the development trend of the preset event.

Preferably, in order to ensure data security, the time series data of the preset event may be stored in the blockchain.

For example: Influenza is a serious global public health problem. Using the time series analysis model to analyze historical flu epidemic information, you can get flu intensity information next week. At the same time, the flu information obtained will not have lag or large errors. problem.

As shown in FIG. 3, it is a schematic diagram of modules of the time series data hysteresis processing device of the present application.

The time series data hysteresis processing apparatus 100 described in this application can be installed in an electronic device. According to the implemented functions, the time series data hysteresis processing device 100 may include a differential data generation module 101, a true curve generation module 102, a prediction curve generation module 103, a loss value calculation module 104, a lag calculation module 105, and a standard time series analysis model The screening module 106 and the data analysis module 107. The module described in this application can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can complete fixed functions, and are stored in the memory of the electronic device.

In this embodiment, the functions of each module/unit are as follows:

The differential data generating module 101 is configured to obtain time series sample data, and perform time series differential processing on the time series sample data to obtain a differential data set;

The true curve generation module 102 is configured to perform curve fitting on the difference data set to obtain a true curve;

The prediction curve generation module 103 is configured to analyze the time series sample data through multiple pre-built time series analysis models to obtain multiple time series analysis data, and perform time series difference processing on the multiple time series analysis data to obtain multiple time series analysis data. Differential analytical data set, curve fitting multiple differential analytical data sets to obtain multiple prediction curves;

The loss value calculation module 104 is configured to use a loss function to calculate a plurality of loss values between the predicted curve and the true curve;

The hysteresis calculation module 105 is configured to calculate the hysteresis factor of each prediction curve based on the loss value, and determine the hysteresis of each prediction curve according to the hysteresis factor;

The standard time series analysis model screening module 106 selects a standard time series analysis model from the plurality of time series analysis models according to the hysteresis of each of the prediction curves;

The data analysis module 107 is configured to use the standard time series analysis model to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.

In detail, the specific implementation of each module of the time series data hysteresis processing device 100 is as follows:

Step 1: The differential data generating module 101 obtains time series sample data, and performs time series differential processing on the time series sample data to obtain a differential data set.

In the embodiment of the present application, the time series sample data includes any data with time series characteristics, such as data of epidemic intensity changes over time, data of object temperature changes over time, and so on. In the embodiment of the present application, a java statement with a data calling function may be used to obtain the time series sample data from a database for storing the time series sample data.

Preferably, before the time-series difference processing is performed on the time-series sample data, the difference data generation module 101 further executes:

Judging whether the time series sample data has stationarity;

When the time series sample data has stationarity, do not perform time series difference processing on the time series sample data;

When the time series sample data does not have stationarity, the time series difference processing of the time series sample data is performed.

Preferably, the differential data generating module 101 of the embodiment of the present application can determine whether the time series sample data has stationarity by detecting whether there is a unit root in the time series sample data. Wherein, the unit root refers to the root modulo 1 in the time series sample data, for example: 1, -1, i, and -i are all 4th unit roots.

When there is no unit root in the time series sample data, it means that the time series sample data has stationarity, and when there is a unit root in the time series sample data, it means that the time series sample data does not have stationarity.

Preferably, the differential data generation module 101 of the embodiment of the present application uses a first-order difference to perform a time-series difference on the time-series sample data, wherein the first-order difference formula is as follows:

Δ _yx =y _x+1 -y _x

Wherein, _Δyx is the first-order difference of the object temperature change data with time, y _x+1 is the data at the time x+1, and y _x is the data at the time x.

In the embodiment of the present application, performing time-series difference processing on the time-series sample data can filter out unstable factors in the time-series sample data.

Step 2: The true curve generation module 102 performs curve fitting on the difference data set to obtain a true curve.

Specifically, the real curve generating module 102 is used to:

Extracting the timing characteristics of the differential data in the differential data set;

Use the time sequence feature to perform coordinate encoding on the differential data in the differential data set;

Mapping the differential data in the differential data set to a pre-built coordinate system by using the coordinate encoding;

The difference data in the coordinate system is connected according to the time sequence feature to obtain the true curve. The true curve is a stable curve, which can represent the curve of the difference data with time.

In the embodiment of the present application, the time sequence feature refers to the time sequence of the differential data; the coordinate encoding of the differential data in the differential data set according to the time sequence feature includes the chronological sequence of the differential data. The difference data is expressed in a preset coordinate system.

For example, establish a coordinate system with time t and t+1 as the horizontal axis, and L _t and L _t+1 as the vertical axis to establish a coordinate system; In the coordinate system, the difference data in the coordinate system is further connected according to the time sequence feature to obtain the true curve.

Step 3. The prediction curve generation module 103 analyzes the time series sample data through multiple pre-built time series analysis models to obtain multiple time series analysis data, and performs time series difference processing on the multiple time series analysis data to obtain multiple time series analysis data. Differential analytical data set, curve fitting is performed on multiple differential analytical data sets to obtain multiple prediction curves.

In the embodiment of the present application, the time sequence analysis model is:

S _t+1 =X _t +Y _t

X _t =αS _t +(1-α)(S _t-1 -Y _t-1 )

Y _t ＝β(X _t -X _t-1 )+(1-β)Y _t-1

Among them, _St represents the difference data of the difference data concentrated at time t, _St-1 represents the difference data of the difference data concentrated at time t-1, X _t represents the smooth value at time t, and X _t-1 represents The smoothed value at time t-1, Y _t represents the trend value at time t, Y _t-1 represents the trend value at time t-1, and S _t+1 represents that the differential analysis data is concentrated in the differential analysis data at time t+1 , Α and β are preset different smoothing coefficients.

The prediction curve generation module 103 described in the embodiment of the present application obtains multiple time series analysis models by customizing different smoothing coefficients α and β.

Specifically, the prediction curve generation module 103 of the embodiment of the present application uses the multiple time series analysis models to analyze the time series sample data to obtain multiple time series analysis data. For example, for the data that the object temperature changes with time as the time series sample data: [(1,2), (2,4), (3,5), (4,8)], where the first digit is the time ID, used to indicate the time of the time series sample data, the second digit is a data ID, used to indicate the numerical value of the object temperature that changes with time; using the time series analysis model to compare the object temperature with time The changed data is analyzed, and multiple time series analysis data are obtained: [(5,9), (6,11), (7,14), (8,16)].

Further, since the obtained time series analysis data may contain unstable factors, the prediction curve generation module 103 described in this embodiment of the present application further performs time series difference processing on the time series analysis data to convert the time series analysis data Unsteady factors in filter out.

In detail, the embodiment of the present application performs the time-series differential processing performed by the differential data generating module 101 on the multiple time-series analysis data to obtain multiple differential analysis data sets.

In the embodiment of the present application, curve fitting is performed on multiple differential analysis data sets to obtain multiple prediction curves.

Preferably, the method of performing curve fitting on multiple differential analysis data sets is the same as the method of performing curve fitting on differential data sets executed by the real curve generating module 102 described above, and will not be repeated here.

Step 4: The loss value calculation module 104 uses a loss function to calculate loss values between multiple predicted curves and the true curve.

Preferably, the loss value calculation module 104 in the embodiment of the present application uses the following formula to calculate the initial loss value between the multiple predicted curves and the true curve:

in,

Represents the initial loss value,

It represents the predicted value of the prediction curve, σ represents the number of the time series analysis models, Y represents the true value of the true curve under different translation degrees, and α represents the error factor.

Further, the loss value calculation module 104 calculates the root mean square error of the initial loss value according to the initial loss value to obtain the loss value, including:

in,

Is the root mean square error of the initial loss value, that is, the loss value,

Represents the initial loss value, and m represents the number of time series analysis models.

Step 5. The hysteresis calculation module 105 calculates the hysteresis factors of multiple prediction curves based on the loss value, and determines the hysteresis of the multiple prediction curves according to the hysteresis factors.

In the embodiment of the present application, the hysteresis calculation module 105 uses the following algorithm to calculate the hysteresis factor τ _m :

Among them, arg min is the set of all arguments that make the function obtain its minimum value,

It is an analytical value, τ is a loss value, i is the starting value of the prediction curve, and n is the end value of the prediction curve.

Further, the hysteresis calculation module 105 of the embodiment of the present application determines the hysteresis of each prediction curve according to the hysteresis factor, including:

Use the following algorithm to calculate the lag score _delay :

Wherein, τ _m is the lag factor of the prediction curve;

It is the smallest root mean square error after _{τ m translation.}

The hysteresis calculation module 105 described in the embodiment of the present application substitutes the hysteresis factors of multiple prediction curves into the above algorithm to obtain the hysteresis of the multiple prediction curves. The hysteresis represents the lag of the prediction curve compared with the true curve degree.

Step 6. The standard time series analysis model screening module 106 selects a standard time series analysis model from the plurality of time series analysis models according to the hysteresis of each of the prediction curves.

In the embodiment of the present application, the standard time series analysis model screening module 106 selects standard time series analysis models from the multiple time series analysis models through the following operations:

The multiple time series analysis models are screened according to the hysteresis, and a standard time series analysis model is obtained.

In the embodiment of the present application, models with hysteresis less than a preset threshold among multiple time series analysis models are screened out to obtain a standard time series analysis model.

The standard time series analysis model is a model selected after comparing the lag and error degree between different time series analysis models, which can solve the long-term lag problem in time series forecasting. The lag is used to measure the degree of lag If the lag of the multiple time series analysis models is smaller, the data processed by the standard time series analysis model will be relatively accurate, and there will be fewer lag problems.

In life, influenza viruses are highly variable and uncertain. Correctly predicting the peak of influenza epidemics is critical to preventing and controlling influenza. The selected standard time series analysis model can solve the problem of lagging forecast results and large forecast errors when forecasting influenza peaks, and improve the accuracy of influenza forecasting.

Step 7. The data analysis module 107 uses the standard time series analysis model to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.

In the embodiment of the present application, the data analysis module 107 analyzes the time sequence to be processed according to the standard time sequence analysis model obtained by screening, and obtains the development trend of the preset event.

Preferably, in order to ensure data security, the time series data of the preset event may be stored in the blockchain.

For example: Influenza is a serious global public health problem. Using the time series analysis model to analyze historical flu epidemic information, you can get flu intensity information next week. At the same time, the flu information obtained will not have lag or large errors. problem.

As shown in FIG. 4, it is a schematic structural diagram of an electronic device implementing the method for processing time series data hysteresis according to the present application.

The electronic device 1 may include a processor 10, a memory 11, and a bus, and may also include a computer program stored in the memory 11 and running on the processor 10, such as a time series data hysteresis processing program 12.

Wherein, the memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (such as SD or DX memory, etc.), magnetic memory, magnetic disk, CD etc. The memory 11 may be an internal storage unit of the electronic device 1 in some embodiments, for example, a mobile hard disk of the electronic device 1. In other embodiments, the memory 11 may also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart media card (SMC), and a secure digital (Secure Digital) equipped on the electronic device 1. , SD) card, flash card (Flash Card), etc. Further, the memory 11 may also include both an internal storage unit of the electronic device 1 and an external storage device. The memory 11 can be used not only to store application software and various data installed in the electronic device 1, such as the code of the time series data hysteresis processing program 12, etc., but also to temporarily store data that has been output or will be output.

The processor 10 may be composed of integrated circuits in some embodiments, for example, may be composed of a single packaged integrated circuit, or may be composed of multiple integrated circuits with the same function or different functions, including one or more Combinations of central processing unit (CPU), microprocessor, digital processing chip, graphics processor, and various control chips, etc. The processor 10 is the control unit of the electronic device, which uses various interfaces and lines to connect the various components of the entire electronic device, and runs or executes programs or modules stored in the memory 11 (for example, executing Time series data hysteresis processing program, etc.), and call the data stored in the memory 11 to execute various functions of the electronic device 1 and process data.

The bus may be a peripheral component interconnect standard (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus can be divided into address bus, data bus, control bus and so on. The bus is configured to implement connection and communication between the memory 11 and at least one processor 10 and the like.

FIG. 4 only shows an electronic device with components. Those skilled in the art can understand that the structure shown in FIG. 3 does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure. Components, or combinations of certain components, or different component arrangements.

For example, although not shown, the electronic device 1 may also include a power source (such as a battery) for supplying power to various components. Preferably, the power source may be logically connected to the at least one processor 10 through a power management device, thereby controlling power The device implements functions such as charge management, discharge management, and power consumption management. The power supply may also include any components such as one or more DC or AC power supplies, recharging devices, power failure detection circuits, power converters or inverters, and power status indicators. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

Further, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and/or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is usually used in the electronic device 1 Establish a communication connection with other electronic devices.

Optionally, the electronic device 1 may also include a user interface. The user interface may be a display (Display) and an input unit (such as a keyboard (Keyboard)). Optionally, the user interface may also be a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, etc. Among them, the display can also be appropriately called a display screen or a display unit, which is used to display the information processed in the electronic device 1 and to display a visualized user interface.

It should be understood that the embodiments are only for illustrative purposes, and are not limited by this structure in the scope of the patent application.

The time series data hysteresis processing program 12 stored in the memory 11 in the electronic device 1 is a combination of multiple instructions. When running in the processor 10, it can realize:

Acquiring time series sample data, and performing time series differential processing on the time series sample data to obtain a differential data set;

Curve fitting the difference data set to obtain a true curve;

Through pre-built multiple time series analysis models, analyze the time series sample data to obtain multiple time series analysis data, perform time series differential processing on the multiple time series analysis data, obtain multiple differential analysis data sets, and analyze multiple differential analysis data. Perform curve fitting on the data set to obtain multiple prediction curves;

Using a loss function to calculate a plurality of loss values between the predicted curve and the real curve;

Calculating the lag factor of each prediction curve based on the loss value, and determining the lag of each prediction curve according to the lag factor;

Filtering out a standard time series analysis model from the plurality of time series analysis models according to the lag of each of the prediction curves;

The standard time series analysis model is used to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.

Further, if the integrated module/unit of the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. It can be non-volatile or volatile. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory) .

Further, the computer usable storage medium may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required by at least one function, etc.; the storage data area may store a block chain node Use the created data, etc.

In the several embodiments provided in this application, it should be understood that the disclosed equipment, device, and method may be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the objectives of the solutions of the embodiments.

In addition, the functional modules in the various embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware, or may be implemented in the form of hardware plus software functional modules.

For those skilled in the art, it is obvious that the present application is not limited to the details of the foregoing exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or basic characteristics of the present application.

Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of this application is defined by the appended claims rather than the above description, and therefore it is intended to fall into the claims. All changes in the meaning and scope of the equivalent elements of are included in this application. Any accompanying diagrams in the claims should not be regarded as limiting the claims involved.

In addition, it is obvious that the word "including" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claims can also be implemented by one unit or device through software or hardware. The second class words are used to indicate names, and do not indicate any specific order.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the application and not to limit them. Although the application has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the application can be Make modifications or equivalent replacements without departing from the spirit and scope of the technical solution of the present application.

Claims (20)

1. A method for processing time series data hysteresis, wherein the method includes:

   Acquiring time series sample data, and performing time series differential processing on the time series sample data to obtain a differential data set;

   Curve fitting the difference data set to obtain a true curve;

   Through pre-built multiple time series analysis models, analyze the time series sample data to obtain multiple time series analysis data, perform time series differential processing on the multiple time series analysis data, obtain multiple differential analysis data sets, and analyze multiple differential analysis data. Perform curve fitting on the data set to obtain multiple prediction curves;

   Using a loss function to calculate a plurality of loss values between the predicted curve and the real curve;

   Calculating the lag factor of each prediction curve based on the loss value, and determining the lag of each prediction curve according to the lag factor;

   Filtering out a standard time series analysis model from the plurality of time series analysis models according to the lag of each of the prediction curves;

   The standard time series analysis model is used to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.
2. 5. The method for processing time series data hysteresis according to claim 1, wherein before the time series differential processing is performed on the time series sample data, the method further comprises:

   Judging whether the time series sample data has stationarity;

   When the time series sample data has stationarity, do not perform time series difference processing on the time series sample data;

   When the time series sample data does not have stationarity, the time series difference processing of the time series sample data is performed.
3. 3. The method for processing time series data hysteresis according to claim 1 or 2, wherein the performing time series difference processing on the time series sample data comprises:

   Use the following formula to perform time-series differential processing on the time-series sample data:

   Δ _yx =y _x+1 -y _x

   Wherein, _Δyx is the first-order difference of the object temperature change data with time, y _x+1 is the data at the time x+1, and y _x is the data at the time x.
4. The method for processing time series data lagging according to claim 1, wherein said performing curve fitting on said differential data set to obtain a true curve comprises:

   Extracting the timing characteristics of the differential data in the differential data set;

   Use the time sequence feature to perform coordinate encoding on the differential data in the differential data set;

   Mapping the differential data in the differential data set to a pre-built coordinate system by using the coordinate encoding;

   The difference data in the coordinate system is connected according to the time sequence feature to obtain the true curve.
5. 3. The method for processing time series data lagging according to claim 1, wherein the time series analysis model is:

   S _t+1 =X _t +Y _t

   X _t =αS _t +(1-α)(S _t-1 -Y _t-1 )

   Y _t ＝β(X _t -X _t-1 )+(1-β)Y _t-1

   Among them, _St represents the difference data of the difference data concentrated at time t, _St-1 represents the difference data of the difference data concentrated at time t-1, X _t represents the smooth value at time t, and X _t-1 represents The smoothed value at time t-1, Y _t represents the trend value at time t, Y _t-1 represents the trend value at time t-1, and S _t+1 represents that the differential analysis data is concentrated in the differential analysis data at time t+1 , Α and β are preset different smoothing coefficients.
6. The method for processing time series data lagging according to claim 1, wherein the calculating the lag factor of the prediction curve based on the loss value comprises:

   _{The lag factor τ m} is calculated using the following algorithm:

   

   Among them, arg min is the set of all arguments that make the function obtain its minimum value,

   

   It is an analytical value, τ is a loss value, i is the starting value of the prediction curve, and n is the end value of the prediction curve.
7. 8. The method for processing time series data lagging according to claim 6, wherein said determining the lagging degree of each said prediction curve according to said lagging factor comprises:

   Use the following algorithm to calculate the lag score _delay :

   

   Wherein, τ _m is the lag factor of the prediction curve;

   

   It is the smallest root mean square error after _{τ m translation.}
8. A method and device for processing time series data hysteresis, wherein the device includes:

   A differential data generation module, used to obtain time series sample data, and perform time series differential processing on the time series sample data to obtain a differential data set;

   The true curve generation module is used to perform curve fitting on the difference data set to obtain a true curve;

   The prediction curve generation module is used to analyze the time series sample data through multiple pre-built time series analysis models to obtain multiple time series analysis data, and perform time series differential processing on the multiple time series analysis data to obtain multiple differential analysis data Set, curve fitting multiple differential analysis data sets to obtain multiple prediction curves;

   A loss value calculation module, configured to use a loss function to calculate a plurality of loss values between the predicted curve and the true curve;

   A hysteresis calculation module, configured to calculate the hysteresis factor of each prediction curve based on the loss value, and determine the hysteresis of each prediction curve according to the hysteresis factor;

   The standard time series analysis model screening module, according to the hysteresis of each of the prediction curves, selects standard time series analysis models from the plurality of time series analysis models;

   The data analysis module is configured to use the standard time series analysis model to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.
9. An electronic device, wherein the electronic device includes:

   At least one processor; and,

   A memory communicatively connected with the at least one processor; wherein,

   The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor can execute the following steps:

   Acquiring time series sample data, and performing time series differential processing on the time series sample data to obtain a differential data set;

   Curve fitting the difference data set to obtain a true curve;

   Through pre-built multiple time series analysis models, analyze the time series sample data to obtain multiple time series analysis data, perform time series differential processing on the multiple time series analysis data, obtain multiple differential analysis data sets, and analyze multiple differential analysis data. Perform curve fitting on the data set to obtain multiple prediction curves;

   Using a loss function to calculate a plurality of loss values between the predicted curve and the real curve;

   Calculating the lag factor of each prediction curve based on the loss value, and determining the lag of each prediction curve according to the lag factor;

   Filtering out a standard time series analysis model from the plurality of time series analysis models according to the lag of each of the prediction curves;

   The standard time series analysis model is used to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.
10. 9. The electronic device according to claim 9, wherein, before the time-series difference processing is performed on the time-series sample data, the method further comprises:

    Judging whether the time series sample data has stationarity;

    When the time series sample data has stationarity, do not perform time series difference processing on the time series sample data;

    When the time series sample data does not have stationarity, the time series difference processing of the time series sample data is performed.
11. The electronic device according to claim 9 or 10, wherein the performing time-series difference processing on the time-series sample data comprises:

    Use the following formula to perform time-series differential processing on the time-series sample data:

    Δ _yx =y _x+1 -y _x

    Wherein, _Δyx is the first-order difference of the object temperature change data with time, y _x+1 is the data at the time x+1, and y _x is the data at the time x.
12. 9. The electronic device according to claim 9, wherein said performing curve fitting on said difference data set to obtain a true curve comprises:

    Extracting the timing characteristics of the differential data in the differential data set;

    Use the time sequence feature to perform coordinate encoding on the differential data in the differential data set;

    Mapping the differential data in the differential data set to a pre-built coordinate system by using the coordinate encoding;

    The difference data in the coordinate system is connected according to the time sequence feature to obtain the true curve.
13. 9. The electronic device of claim 9, wherein the timing analysis model is:

    S _t+1 =X _t +Y _t

    X _t =αS _t +(1-α)(S _t-1 -Y _t-1 )

    Y _t ＝β(X _t -X _t-1 )+(1-β)Y _t-1

    Among them, _St represents the differential data of the differential data concentrated at time t, _St-1 represents the differential data of the differential data concentrated at time t-1, X _t represents the smooth value at time t, and X _t-1 represents The smoothed value at time t-1, Y _t represents the trend value at time t, Y _t-1 represents the trend value at time t-1, and S _t+1 represents that the differential analysis data is concentrated in the differential analysis data at time t+1 , Α and β are preset different smoothing coefficients.
14. 9. The electronic device according to claim 9, wherein the calculating the lag factor of the prediction curve based on the loss value comprises:

    _{The lag factor τ m} is calculated using the following algorithm:

    

    Among them, arg min is the set of all arguments that make the function obtain its minimum value,

    

    It is an analytical value, τ is a loss value, i is the starting value of the prediction curve, and n is the end value of the prediction curve.
15. 9. The electronic device according to claim 9, wherein said determining the hysteresis of each said prediction curve according to said hysteresis factor comprises:

    Use the following algorithm to calculate the lag score _delay :

    

    Wherein, τ _m is the lag factor of the prediction curve;

    

    It is the smallest root mean square error after _{τ m translation.}
16. A computer-readable storage medium storing a computer program, wherein the computer program is executed by a processor to implement the following steps:

    Acquiring time series sample data, and performing time series differential processing on the time series sample data to obtain a differential data set;

    Curve fitting the difference data set to obtain a true curve;

    Through pre-built multiple time series analysis models, analyze the time series sample data to obtain multiple time series analysis data, perform time series differential processing on the multiple time series analysis data, obtain multiple differential analysis data sets, and analyze multiple differential analysis data. Perform curve fitting on the data set to obtain multiple prediction curves;

    Using a loss function to calculate a plurality of loss values between the predicted curve and the real curve;

    Calculating the lag factor of each prediction curve based on the loss value, and determining the lag of each prediction curve according to the lag factor;

    Filtering out a standard time series analysis model from the plurality of time series analysis models according to the lag of each of the prediction curves;

    The standard time series analysis model is used to perform data analysis on the time series data of the preset event input by the user to obtain the development trend of the preset event.
17. 15. The computer-readable storage medium according to claim 16, wherein, before the time-series difference processing is performed on the time-series sample data, the method further comprises:

    Judging whether the time series sample data has stationarity;

    When the time series sample data has stationarity, do not perform time series difference processing on the time series sample data;

    When the time series sample data does not have stationarity, the time series difference processing of the time series sample data is performed.
18. 18. The computer-readable storage medium according to claim 16 or 17, wherein the performing time-series difference processing on the time-series sample data comprises:

    Use the following formula to perform time-series differential processing on the time-series sample data:

    Δ _yx =y _x+1 -y _x

    Wherein, _Δyx is the first-order difference of the object temperature change data with time, y _x+1 is the data at the time x+1, and y _x is the data at the time x.
19. 15. The computer-readable storage medium of claim 16, wherein said performing curve fitting on the difference data set to obtain a true curve comprises:

    Extracting the timing characteristics of the differential data in the differential data set;

    Use the time sequence feature to perform coordinate encoding on the differential data in the differential data set;

    Mapping the differential data in the differential data set to a pre-built coordinate system by using the coordinate encoding;

    The difference data in the coordinate system is connected according to the time sequence feature to obtain the true curve.
20. 15. The computer-readable storage medium of claim 16, wherein the timing analysis model is:

    S _t+1 =X _t +Y _t

    X _t =αS _t +(1-α)(S _t-1 -Y _t-1 )

    Y _t ＝β(X _t -X _t-1 )+(1-β)Y _t-1

    Among them, _St represents the difference data of the difference data concentrated at time t, _St-1 represents the difference data of the difference data concentrated at time t-1, X _t represents the smooth value at time t, and X _t-1 represents The smoothed value at time t-1, Y _t represents the trend value at time t, Y _t-1 represents the trend value at time t-1, and S _t+1 represents that the differential analysis data is concentrated in the differential analysis data at time t+1 , Α and β are preset different smoothing coefficients.

Images