WO2023197502A1

WO2023197502A1 - Comprehensive power evaluation method and apparatus

Info

Publication number: WO2023197502A1
Application number: PCT/CN2022/115190
Authority: WO
Inventors: 程敏; 林溪桥; 陈志君; 覃惠玲; 卢纯颢; 王鹏; 周春丽
Original assignee: 广西电网有限责任公司
Priority date: 2022-04-11
Filing date: 2022-08-26
Publication date: 2023-10-19
Also published as: CN114792196A

Abstract

The present application relates to a comprehensive power evaluation method and apparatus. The method in the present application comprises: acquiring a power indicator and an economic indicator; inputting the power indicator and the economic indicator into an indicator pool, so as to execute a preprocessing operation, wherein the preprocessing operation comprises missing-value adjustment, seasonal factor adjustment, indicator and GDP correlation analysis, and indicator normalization processing; performing regression on a GDP and preprocessed indicator data in the indicator pool by using a LASSO regression method; and determining, on the basis of a regression coefficient generated by means of regression, an evaluation index corresponding to the power indicator and the economic indicator. In the method, variable selection and weighting are performed on preprocessed indicators, correlation analysis is performed on the indicators and a GDP, variables having a relatively high correlation are retained and relatively high weights are assigned thereto, and variables having a relatively low correlation are discarded. In this way, comprehensive power evaluation index data obtained by fully utilizing information contained in the GDP itself is relatively accurate, and comprehensive power evaluation is realized.

Description

An electric power comprehensive evaluation method and device

Technical field

The present application relates to the technical field of electric power evaluation, and in particular to a comprehensive electric power evaluation method and device.

Background technique

The power comprehensive index is a comprehensive index that integrates power big data and contains a large amount of information about the power industry and social and economic development. The power comprehensive index can be used to comprehensively evaluate the development of the power economy. At present, there are mainly three types of research on comprehensive indexes. One is to use a representative single indicator in the field of economics, such as total factor productivity, labor productivity, etc. to directly measure economic development; the other is to build a comprehensive index evaluation system to measure economic development. development; the third is to construct a prosperity index as a qualitative indicator to reflect the status or development trend of a specific survey group or a certain socio-economic phenomenon.

In the existing technology, the method of constructing a comprehensive evaluation system is generally used to conduct comprehensive evaluation of electric power. However, the existing evaluation method has problems such as imperfect evaluation index pool, complicated final index system, and failure to fully utilize the information contained in GDP. Among them, the indicators are incomplete. It means that because there is no accurate definition of measurement indicators, there are relatively large differences in the indicator pool. Some indicators have not been considered, some indicators do not meet the actual needs of a certain region, and some indicators do not meet the requirements of the development of the times, especially There is a relative lack of research on the quality assessment of regional economic development involving the application of power big data; the complexity of the final indicator system means that all indicators in the evaluation system are included in the synthesis of the final index, so that the weight of each indicator will be further weakened And the change in weight increases, making the final indicator system complicated; not fully utilizing the information contained in GDP refers to defining GDP as an indicator rather than a dependent variable containing information. The weight of the index GDP obtained in this way is generally larger. is small, and GDP has a greater correlation with the existing index system. Due to the imperfect evaluation index pool, the complexity of the final index system, and the failure to fully utilize the information contained in GDP, the obtained power comprehensive evaluation index data is inaccurate and cannot produce actual evaluation significance.

Contents of the invention

The embodiments of this application provide a power comprehensive evaluation method and device, which are used to conduct supervised learning and high-dimensional data analysis through GDP year-on-year growth rate, and make full use of the information contained in GDP itself to more accurately calculate the power comprehensive evaluation index.

The first aspect of the embodiments of this application provides a comprehensive evaluation method for electric power, including:

Get electricity indicators and economic indicators;

Input the power indicator and the economic indicator into the indicator pool to perform preprocessing operations. The preprocessing operations include missing value adjustment, seasonal factor adjustment, correlation analysis between indicators and GDP, and indicator normalization processing;

Regress GDP and the preprocessed indicator data in the indicator pool through the Least absolute shrinkage and selection operator regression method;

The evaluation index corresponding to the electric power indicator and the economic indicator is determined based on the regression coefficient generated by the regression.

Optionally, the input of the power indicator and the economic indicator into the indicator pool to adjust missing values of the indicators includes:

A small amount of missing data in the indicator pool is supplemented through smooth interpolation;

Interpolate data with a large amount of missing data in the indicator pool according to the data proportion analysis method.

Optionally, inputting the power indicator and the economic indicator into the indicator pool to adjust the indicator seasonal factors includes:

The year-on-year adjustment method and the X-12-ARIMA seasonal adjustment method are used to remove seasonal factors from the indicator data in the indicator pool.

Optionally, the power indicator and the economic indicator are input into an indicator pool to perform correlation analysis between indicators and GDP, including:

Conduct correlation analysis between the year-on-year growth rate of indicators in the indicator pool and the year-on-year growth rate of GDP.

Optionally, inputting the power indicator and the economic indicator into an indicator pool to normalize the indicators includes:

The year-on-year growth rate in the indicator pool is normalized, and the normalization formula is:

Among them, max( _xi ) represents the maximum value of the sequence { _xi }; min( _xi ) represents the minimum value of the sequence { _xi }.

Optionally, the formula of the LASSO regression is

in

is the preprocessed GDP year-on-year growth rate,

is the preprocessed indicator pool,

is the preprocessed indicator, β is the regression coefficient vector, and ω _j is the penalty weight of each indicator.

Optionally, determining the evaluation index corresponding to the power indicator and the economic indicator based on the regression coefficient generated by the regression includes:

Determine the weight of the electric power indicator and the economic indicator based on the regression coefficient generated by the regression;

The electric power indicator and the economic indicator are assigned and calculated according to the weight to obtain an evaluation index. The calculation formula is:

where β is the corresponding parameter vector.

The second aspect of the embodiment of the present application provides a comprehensive power evaluation device, including:

Acquisition unit, used to obtain power indicators and economic indicators;

An execution unit is used to input the power indicator and the economic indicator into the indicator pool to perform preprocessing operations. The preprocessing operations include adjusting the missing values of the indicators, adjusting the seasonal factors contained in the indicators, performing correlation analysis on the indicators, and The indicators are normalized;

A regression unit used to regress GDP and the preprocessed indicator data in the indicator pool through the LASSO regression method;

A determining unit configured to determine an evaluation index corresponding to the electric power index and the economic index based on the regression coefficient generated by the regression.

Optionally, the execution unit includes a missing value adjustment module, a seasonal factor adjustment module, an indicator and GDP correlation analysis module, and an indicator normalization processing module:

The missing value adjustment module is used to supplement a small amount of missing data in the indicator pool through smooth interpolation method and to interpolate a large amount of missing data in the indicator pool according to the data proportion analysis method;

The seasonal factor adjustment module is used to remove seasonal factors from the indicator data in the indicator pool using the year-on-year adjustment method and the X-12-ARIMA seasonal adjustment method;

The indicator and GDP correlation analysis module is used to perform correlation analysis on the year-on-year growth rate of indicator data in the indicator pool and the year-on-year growth rate of GDP;

The indicator normalization processing module is used to normalize the year-on-year growth rate in the indicator pool.

Optionally, the determination unit includes a determination module and a calculation module:

A determination module, configured to determine the weight of the power indicator and the economic indicator based on the regression coefficient generated by the regression;

The calculation module is used to assign and calculate the power indicator and the economic indicator according to the weight to obtain an evaluation index. The calculation formula is:

where β is the corresponding parameter vector.

The third aspect of the embodiment of the present application provides a comprehensive power evaluation device, including:

Processor, memory, input/output unit, bus;

The processor is connected to the memory, the input and output unit and the bus;

The processor specifically performs the following operations:

Get electricity indicators and economic indicators;

Optionally, the processor is also configured to perform operations of any optional solution in the first aspect.

The fourth aspect of the embodiment of the present application provides a computer-readable storage medium for a comprehensive power evaluation method, including:

The computer-readable storage medium stores a program, and the program executes the aforementioned power comprehensive evaluation method on the computer.

It can be seen from the above technical solutions that the embodiments of the present application have the following advantages: In this application, after obtaining the power indicators and economic indicators, the system inputs the power indicators and economic indicators into the indicator pool to adjust missing values, seasonal factors, and indicators. Pre-processing such as correlation analysis with GDP and indicator normalization processing, and then using the LASSO regression method to regress GDP and the pre-processed indicator data in the indicator pool, and finally determine the corresponding evaluation of power indicators and economic indicators based on the regression coefficients generated by the regression. Index, this method uses GDP year-on-year growth rate for supervised learning and high-dimensional data analysis, and performs variable selection and weighting on the indicators. Variables with greater correlation analysis with GDP will be retained and given higher weights, and variables with greater correlation with GDP will be retained and given higher weights. Small variables are discarded, so that the comprehensive evaluation index data for electricity obtained by making full use of the information contained in GDP itself is more accurate and achieves a comprehensive evaluation of electricity.

Description of the drawings

Figure 1 is a schematic flow chart of an embodiment of the power comprehensive evaluation method in the embodiment of the present application;

Figure 2 is a schematic flow diagram of another embodiment of the power comprehensive evaluation method in the embodiment of the present application;

Figure 3 is a schematic structural diagram of an embodiment of the power comprehensive evaluation device in the embodiment of the present application;

Figure 4 is a schematic structural diagram of another embodiment of the power comprehensive evaluation device in the embodiment of the present application.

Detailed ways

In this embodiment, the power comprehensive evaluation method can be implemented in the system, on the server, or on the terminal, and is not specifically limited.

Please refer to Figure 1. The embodiment of the present application is described using a system example. An example of the power comprehensive evaluation method in the embodiment of the present application includes:

101. The system obtains power indicators and economic indicators;

In this embodiment, the system uses power consumption indicators and economic development indicators as elements of comprehensive evaluation, and the introduction of power indicators and economic indicators can solve the problem of incomplete evaluation indicator pools. The definitions of indicators at all levels are based on the specificity of each subject. conduct.

For example, power indicators can include power supply, power purchase, power sales, line loss power, line loss rate, network loss rate, hydropower generation, thermal power generation, oil power generation, coal power generation, gas power generation, Photovoltaic power generation, installed capacity, total profit of the power company, net profit of the power company, economic added value of the power company, return on net assets of the power company, preservation and appreciation rate of state-owned capital of the power company, return on total assets of the power company, operating income of the power company, Asset-liability ratio of power companies, unit power supply cost, unit power purchase cost, primary industry electricity consumption, secondary industry electricity consumption, tertiary industry electricity consumption, urban and rural residential electricity consumption, central urban area residential comprehensive voltage qualification rate, Comprehensive voltage qualification rate for urban residents, comprehensive voltage qualification rate for rural residents, third-party customer satisfaction, medium-voltage line failure rate, average number of power outages for power outage users, smart meter coverage, and electricity consumption in key industries in the region, etc.;

Economic indicators can include crude oil processing output, output of ten non-ferrous metals, soda ash output, chemical fertilizer output, ethylene output, cement output, crude steel output, alumina output, pig iron output, coke output, raw coal output, added value of the real estate industry, social Total retail sales of consumer goods, freight volume, total import and export value, per capita consumption expenditure of urban households, per capita food consumption expenditure of urban households, per capita residential consumption expenditure of urban households, per capita consumption expenditure of education, culture and entertainment services of urban households, Consumer confidence index, per capita disposable income of urban households, commercial housing sales prices, consumer price classification index, industrial producer purchasing price index, industrial producer ex-factory price index, fixed asset investment completion, total inventory, railway manufacturing industry Monthly freight volume of finished products and monthly customs exports in RMB, etc. In addition, it should be noted that there are no restrictions on the names of indicators at all levels.

102. The system inputs power indicators and economic indicators into the indicator pool to perform preprocessing operations. The preprocessing operations include missing value adjustment, seasonal factor adjustment, correlation analysis between indicators and GDP, and indicator normalization processing;

After the system selects power indicators and economic indicators and inputs them into the indicator pool, it preprocesses the indicator data in the indicator pool. The preprocessing includes missing value adjustment, that is, smoothing interpolation is used to supplement a small amount of missing data and perform proportional analysis on a large amount of missing data; Seasonal factor adjustment is to remove the influence of seasonal factors in the index model; indicator and GDP correlation analysis is to conduct correlation analysis between year-on-year growth rate and GDP year-on-year growth rate; indicator normalization processing is based on positive indicators and negative indicators. The year-on-year growth rate is normalized to the indicator.

103. The system uses the LASSO regression method to regress GDP and the preprocessed indicator data in the indicator pool;

The system builds a model through the LASSO regression method, and then performs variable selection, in which variable selection is achieved by adjusting the parameters in the model. Specifically, the larger the parameter, the greater the penalty for a linear model with more variables, thus obtaining a model with more variables. Fewer models. The determination of the model is adjusted through the cross-validation method. For example, given a parameter λ, substitute λ into the model for cross-validation, select the λ value with the smallest cross-validation error, and then re-fit it with all the data based on the optimal λ value obtained. Just fit the model.

The expression of the LASSO model in this embodiment is as follows:

in,

is the year-on-year vector for GDP,

is the preprocessed indicator matrix, and ω _j is the weight of the penalty item for a certain indicator. If the weight of the indicator is larger and the indicator pool is selected more times, the penalty weight should be smaller. After solving the model, the variables corresponding to the estimated parameter β _j being zero can be directly excluded from the composition of the indicator pool in the final index.

104. Based on the regression coefficient system generated by regression, the evaluation index corresponding to the power indicator and economic indicator is determined.

For the d indicators selected through the above steps, the corresponding parameter vector is β. Then the final index of the i-th evaluation individual at time t can be calculated by the following formula:

The system uses this calculation method to make the calculation process faster. The obtained power comprehensive evaluation index structure has the characteristics of continuity, order and low variance. It uses supervised learning and high-dimensional data analysis to reduce the complexity of the variable system in index construction, thereby reducing the complexity of the variable system. It weakens the problem of the contribution of indicators to the index, and effectively extracts the GDP information in the indicator system, thereby better realizing the comprehensive evaluation of electric power using the electric power economic index.

Please refer to Figure 2. The embodiment of the present application is described using a system example. Another embodiment of the power comprehensive evaluation index method in the embodiment of the present application includes:

201. The system obtains power indicators and economic indicators;

Step 201 in this embodiment is similar to step 101 in the previous embodiment, and will not be described again here.

202. The system inputs the power indicators and economic indicators into the indicator pool to adjust the missing values of the indicators;

The system uses smooth interpolation to supplement a small amount of missing data in the indicator pool. The formula is as follows:

Among them, x _i,t-1 and x _i,t+1 are the values of the i-th indicator t-1 quarterly or monthly and t+1 quarterly or monthly,

That is, the value that needs to be inserted, that is, the value of t quarter or month of the i-th indicator is missing, insert

as an alternative.

For data with a large amount of missing data, based on the ratio analysis of existing data and national data, the value is interpolated by multiplying the national data by the average. The formula is as follows:

n is the total number of quarters,

is the known value of the i-th indicator in quarter k, x _i,k is the national value of the i-th indicator in quarter k, x _i,t is the national value of the i-th indicator in quarter t,

That is the value that needs to be inserted.

203. The system inputs power indicators and economic indicators into the indicator pool to adjust seasonal factors;

The system first uses the year-on-year adjustment method to remove seasonal factors. If there is still an obvious seasonal effect after year-on-year adjustment, the X-12-ARIMA seasonal adjustment method is used. The formula for the year-on-year adjustment method is:

x _i,t is the value of the i-th indicator in quarter t, and x _i,t-12 is the value of the indicator in the same period last year. Subtract all indicator values from the same period last year, and divide the resulting difference by the same period last year x _i,t-12 to obtain the year-on-year growth rate of the indicator. The obtained year-on-year growth rate removes seasonal effects and dimensions. The present invention first adjusts seasonal factors by calculating the year-on-year growth rate, which facilitates calculation and can effectively remove seasonal effects.

The formula of the X-12-ARIMA seasonal adjustment method is:

Y _t =T _t ×S _t ×I _t

T _t represents the trend of the time series, S _t represents the seasonality of the time series, and I _t represents irregular disturbance factors. The X-12-ARIMA seasonal adjustment method decomposes the trend and seasonality of the time series through the moving average algorithm, and then uses The original series is divided by the seasonal effect to obtain a new time series result with no seasonality.

204. The system inputs power indicators and economic indicators into the indicator pool to analyze the correlation between indicators and GDP;

The system performs correlation analysis on each year-on-year growth rate and GDP year-on-year growth rate. Among the indicators, indicators with greater correlation and stable performance in the time dimension are selected into the indicator system; indicators that are positively correlated with GDP year-on-year growth rate are marked as positive. indicator, otherwise it is recorded as an inverse indicator.

205. The system inputs the power indicators and economic indicators into the indicator pool for normalization of the indicators;

In order to facilitate the next weighting step, the system normalizes all year-on-year growth rates.

The normalization formula is as follows:

206. The system uses the LASSO regression method to regress GDP and the preprocessed indicator data in the indicator pool;

Step 206 in this embodiment is similar to step 103 in the previous embodiment, and will not be described again here.

207. The system determines the weight of power indicators and economic indicators based on the regression coefficients generated by regression;

The system uses the regression coefficients generated by the LASSO regression method in the indicator pool as the weight of the indicator system, and empowers the indicators so that value assignment calculations can be performed based on weighted power indicators and economic indicators.

208. The system assigns values to power indicators and economic indicators based on weights and calculates the evaluation index.

The system assigns values to power indicators and economic indicators based on the set weights and finally obtains a more accurate evaluation index. The present invention can use the information contained in GDP to play a supervised learning role through the LASSO regression method, thereby better selecting and empowering variables for high-dimensional electric power big data. Supervised learning and high-dimensional data analysis can effectively reduce the problem of redundant variable systems in index construction, thereby weakening the contribution of indicators to the index, and effectively extract the GDP information in the indicator system, thereby better realizing the use of electric power economic indexes for comprehensive evaluation of electric power.

Please refer to Figure 3. An embodiment of the power comprehensive evaluation device in the embodiment of this application includes:

The acquisition unit 301 is used to acquire power indicators and economic indicators;

The execution unit 302 is used to input power indicators and economic indicators into the indicator pool to perform preprocessing operations. The preprocessing operations include adjusting missing values of indicators, adjusting seasonal factors contained in indicators, performing correlation analysis on indicators, and normalizing indicators. deal with;

The regression unit 303 is used to regress GDP and preprocessed indicator data in the indicator pool through the LASSO regression method;

The determination unit 304 is configured to determine the evaluation index corresponding to the power indicator and the economic indicator based on the regression coefficient generated by the regression.

In this embodiment, the execution unit 302 includes a missing value adjustment module, a seasonal factor adjustment module, an indicator and GDP correlation analysis module, and an indicator normalization processing module.

The missing value adjustment module 3021 is used to supplement a small amount of missing data in the indicator pool through the smooth interpolation method and to interpolate a large amount of missing data in the indicator pool according to the data proportion analysis method.

The seasonal factor adjustment module 3022 is used to remove seasonal factors from the data in the indicator pool using the year-on-year adjustment method and the X-12-ARIMA seasonal adjustment method.

Indicator and GDP correlation analysis 3023 is used to perform correlation analysis between the year-on-year growth rate of data in the indicator pool and the year-on-year growth rate of GDP.

The indicator normalization processing module 3024 is used to normalize the year-on-year growth rate in the indicator pool.

In this embodiment, the determination unit 304 includes a determination module 3041 and a calculation module 3042.

Determination module 3041, used to determine the weight of power indicators and economic indicators based on the regression coefficients generated by regression;

The calculation module 3042 is used to assign and calculate the power indicators and economic indicators according to the weights to obtain the evaluation index.

In this embodiment, after the acquisition unit 301 obtains the electric power index and the economic index, the execution unit 302 inputs the electric power index and the economic index into the index pool for preprocessing operation, in which the missing value adjustment module 3021 uses the smooth interpolation method to adjust the small amount of data in the index pool. Missing data are supplemented and large missing data in the indicator pool are interpolated according to the data proportion analysis method; the seasonal factor adjustment module 3022 uses the year-on-year adjustment method and the X-12-ARIMA seasonal adjustment method to remove seasonal factors from the data in the indicator pool. ; The seasonal factor adjustment module 3022 uses the year-on-year adjustment method and the X-12-ARIMA seasonal adjustment method to remove seasonal factors from the data in the indicator pool; the indicator normalization processing module 3024 normalizes the year-on-year growth rate in the indicator pool, and the regression unit 303 Then the GDP and the preprocessed indicator data in the indicator pool are regressed through the LASSO regression method. The determination module in the determination unit 304 determines the weight according to the regression coefficient generated by the regression. Finally, the calculation module 3042 performs assignment calculation on the power indicators and economic indicators based on the weight. A relatively accurate evaluation index is obtained. The present invention makes full use of the information contained in GDP itself to obtain relatively accurate electric power comprehensive evaluation index data, thereby realizing a comprehensive evaluation of electric power.

Please refer to Figure 4. Another embodiment of the power comprehensive evaluation device in the embodiment of the present application includes:

Processor 401, memory 402, input determination unit 403, bus 404;

The processor 401 is connected to the memory 402, the input determination unit 403 and the bus 404;

Processor 401 performs the following operations:

Get electricity indicators and economic indicators;

Input power indicators and economic indicators into the indicator pool to perform preprocessing operations. The preprocessing operations include missing value adjustment, seasonal factor adjustment, correlation analysis between indicators and GDP, and indicator normalization processing;

Use the LASSO regression method to regress GDP and the preprocessed indicator data in the indicator pool;

The evaluation index corresponding to the power indicator and economic indicator is determined based on the regression coefficient generated by the regression.

Optionally, the functions of the processor 401 correspond to the steps in the aforementioned embodiment shown in FIGS. 1 to 2 , and will not be described again here.

Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be described again here.

In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or can be integrated into another system, or some features can be ignored, or not implemented. On the other hand, the coupling or direct coupling or communication connection between each other shown or discussed may be through some interfaces, and the indirect coupling or communication connection of the devices or units may be in electrical, mechanical or other forms.

The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or they may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

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

If the integrated unit 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. Based on this understanding, the technical solution of the present application is essentially or contributes to the existing technology, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , including several instructions to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM, read-only memory), random access memory (RAM, random access memory), magnetic disk or optical disk and other media that can store program code. .

Claims

A comprehensive evaluation method for electric power, which is characterized by including:

Get electricity indicators and economic indicators;

Input the power indicator and the economic indicator into the indicator pool to perform preprocessing operations. The preprocessing operations include missing value adjustment, seasonal factor adjustment, correlation analysis between indicators and GDP, and indicator normalization processing;

Regress GDP and the preprocessed indicator data in the indicator pool through the Least absolute shrinkage and selection operator regression method;

The evaluation index corresponding to the electric power indicator and the economic indicator is determined based on the regression coefficient generated by the regression.
The method according to claim 1, characterized in that said inputting the power indicator and the economic indicator into the indicator pool to adjust the missing value of the indicator includes:

A small amount of missing data in the indicator pool is supplemented through smooth interpolation;

Interpolate data with a large amount of missing data in the indicator pool according to the data proportion analysis method.
The method according to claim 1, characterized in that, inputting the power indicator and the economic indicator into an indicator pool to adjust the indicator seasonal factors includes:

The year-on-year adjustment method and the X-12-ARIMA seasonal adjustment method are used to remove seasonal factors from the indicator data in the indicator pool.
The method according to claim 1, characterized in that said inputting said electric power indicator and said economic indicator into an indicator pool to perform correlation analysis between indicators and GDP includes:

Conduct correlation analysis between the year-on-year growth rate of indicators in the indicator pool and the year-on-year growth rate of GDP.
The method according to claim 4, characterized in that said inputting the power indicator and the economic indicator into an indicator pool to normalize the indicators includes:

The year-on-year growth rate of the indicators in the indicator pool is normalized. The normalization formula is:
Among them, max( xi ) represents the maximum value of the sequence { xi }; min( xi ) represents the minimum value of the sequence { xi }.
The method according to claim 1, characterized in that the formula of the LASSO regression is
in
is the preprocessed GDP year-on-year growth rate,
is the preprocessed indicator pool,
is the preprocessed indicator, β is the regression coefficient vector, and ω j is the penalty weight of each indicator.
The method of claim 1, wherein determining the evaluation index corresponding to the power indicator and the economic indicator based on the regression coefficient generated by the regression includes:

Determine the weight of the electric power indicator and the economic indicator based on the regression coefficient generated by the regression;

The electric power indicator and the economic indicator are assigned and calculated according to the weight to obtain an evaluation index. The calculation formula is:
where β is the corresponding parameter vector.
An electric power comprehensive evaluation device, characterized by including:

Acquisition unit, used to obtain power indicators and economic indicators;

Execution unit, used to input the power indicator and the economic indicator into the indicator pool to perform preprocessing operations. The preprocessing operations include adjusting missing values of the indicators, adjusting seasonal factors contained in the indicators, performing correlation analysis on the indicators, and The indicators are normalized;

A regression unit used to regress GDP and the preprocessed indicator data in the indicator pool through the LASSO regression method;

A determining unit configured to determine an evaluation index corresponding to the electric power index and the economic index based on the regression coefficient generated by the regression.