Power demand prediction method and device based on industry prosperity index, computer equipment and storage medium

Abstract

The invention relates to an industry prosperity index-based power demand prediction method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring a preset industry index corresponding to a to-be-analyzed industry; performing index screening processing on the preset industry index to obtain a target industry index corresponding to the to-be-analyzed industry; obtaining current index data of the to-be-analyzed industry under the target industry index; inputting the current index data into the trained index data prediction model to obtain prediction index data of the to-be-analyzed industry under the target industry index; according to the prediction index data, determining a prosperity index corresponding to the to-be-analyzed industry; and according to the business index, determining predicted power demand information corresponding to the to-be-analyzed industry. By adopting the method, the prediction accuracy of the power demand can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid, in particular to a power demand prediction method and device based on industry prosperity index, computer equipment, computer readable storage medium and computer program product. BACKGROUND

[0002] At present, in order to optimize the resource allocation of the power system, how to accurately predict the power demand is very important.

[0003] In the traditional technology, when predicting the power demand, time series analysis, machine learning and other methods are generally used. However, the influence of economic situation change on power consumption is easily ignored in the prediction process, which affects the prediction accuracy of power demand. SUMMARY

[0004] Therefore, it is necessary to provide a power demand prediction method and device based on industry prosperity index, computer equipment, computer readable storage medium and computer program product, which can improve the prediction accuracy of power demand.

[0005] In a first aspect, the present application provides a power demand prediction method based on industry prosperity index, comprising:

[0006] In response to a power demand prediction request for an industry to be analyzed, a preset industry index corresponding to the industry to be analyzed is obtained;

[0007] The preset industry index is subjected to index screening processing to obtain a target industry index corresponding to the industry to be analyzed;

[0008] Current index data of the industry to be analyzed under the target industry index is obtained;

[0009] The current index data is input into a trained index data prediction model to obtain predicted index data of the industry to be analyzed under the target industry index;

[0010] According to the predicted index data, a prosperity index corresponding to the industry to be analyzed is determined;

[0011] Through a trained power demand information prediction model, based on the prosperity index, predicted power demand information corresponding to the industry to be analyzed is determined.

[0012] In one embodiment, the preset industry index is subjected to index screening processing to obtain a target industry index corresponding to the industry to be analyzed, comprising:

[0013] A first feature vector of the preset industry index is extracted, and a second feature vector corresponding to the preset industry information of the industry to be analyzed is extracted;

[0014] determine a similarity between the preset industry indicator and the preset industry information according to the first feature vector and the second feature vector;

[0015] select a first candidate industry indicator with a similarity greater than a preset similarity between the preset industry indicator and the preset industry information from the preset industry indicators;

[0016] obtain a data missing rate and an indicator coverage rate corresponding to the first candidate industry indicator, and select a second candidate industry indicator with a data missing rate less than a preset data missing rate and an indicator coverage rate greater than a preset industry indicator coverage rate from the first candidate industry indicators;

[0017] perform an indicator screening process on the second candidate industry indicators to obtain the target industry indicator.

[0018] In one embodiment, the performing of the indicator screening process on the second candidate industry indicators to obtain the target industry indicator comprises:

[0019] obtaining a correlation between the second candidate industry indicators;

[0020] performing an indicator screening process on the second candidate industry indicators according to the correlation between the second candidate industry indicators to obtain third candidate industry indicators; the correlation between the third candidate industry indicators is less than a preset correlation;

[0021] determining a variance inflation factor value between the third candidate industry indicators according to a regression determination coefficient between the third candidate industry indicators;

[0022] performing a re-indicator screening process on the third candidate industry indicators according to the variance inflation factor value between the third candidate industry indicators to obtain fourth candidate industry indicators; the variance inflation factor value between the fourth candidate industry indicators is less than a preset variance inflation factor value;

[0023] obtaining the target industry indicator based on the fourth candidate industry indicators.

[0024] In one embodiment, before inputting the current indicator data into the trained indicator data prediction model to obtain the predicted indicator data of the target industry indicator of the industry to be analyzed, the method further comprises:

[0025] performing a smoothing process on the current indicator data to obtain processed current indicator data; the data frequency of the processed current indicator data is less than the data frequency of the current indicator data;

[0026] The current index data is input into the trained index data prediction model to obtain predicted index data of the industry to be analyzed under the target industry index.

[0027] The processed current index data is input into the trained index data prediction model to obtain predicted index data of the industry to be analyzed under the target industry index.

[0028] In one of the embodiments, the predicted index data is used to determine the corresponding industry prosperity index of the industry to be analyzed, including:

[0029] The predicted index data is normalized according to the historical index data corresponding to the predicted index data to obtain processed predicted index data.

[0030] The scaling factor corresponding to the processed predicted index data is determined according to the historical index data.

[0031] The processed predicted index data and the scaling factor are used to query a preset corresponding relationship to obtain the corresponding industry prosperity index of the industry to be analyzed. The preset corresponding relationship represents the corresponding relationship between the processed predicted index data, the scaling factor, and the industry prosperity index.

[0032] In one of the embodiments, the trained power demand information prediction model is used to determine the corresponding predicted power demand information of the industry to be analyzed based on the industry prosperity index, including:

[0033] The current prosperity interval corresponding to the industry prosperity index is determined.

[0034] The current prosperity interval is input into the trained power demand information prediction model to obtain the predicted probability of the industry to be analyzed under each preset power demand information.

[0035] The preset power demand information with the maximum predicted probability is selected from the preset power demand information as the predicted power demand information corresponding to the industry to be analyzed.

[0036] In a second aspect, the present application further provides a power demand prediction device based on an industry prosperity index, including:

[0037] An index acquisition module is configured to acquire preset industry indexes corresponding to an industry to be analyzed in response to a power demand prediction request for the industry to be analyzed.

[0038] An index screening module is configured to perform index screening processing on the preset industry indexes to obtain target industry indexes corresponding to the industry to be analyzed.

[0039] a data acquisition module, configured to acquire current index data of the to-be-analyzed industry under the target industry index;

[0040] a data prediction module, configured to input the current index data into a trained index data prediction model to obtain predicted index data of the to-be-analyzed industry under the target industry index;

[0041] an index determination module, configured to determine a corresponding industry prosperity index of the to-be-analyzed industry according to the predicted index data;

[0042] a demand prediction module, configured to determine corresponding predicted power demand information of the to-be-analyzed industry based on the industry prosperity index by using a trained power demand information prediction model.

[0043] In a third aspect, the present application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0044] in response to a power demand prediction request for a to-be-analyzed industry, acquiring preset industry indexes corresponding to the to-be-analyzed industry;

[0045] performing index screening processing on the preset industry indexes to obtain target industry indexes corresponding to the to-be-analyzed industry;

[0046] acquiring current index data of the to-be-analyzed industry under the target industry indexes;

[0047] inputting the current index data into a trained index data prediction model to obtain predicted index data of the to-be-analyzed industry under the target industry indexes;

[0048] determining a corresponding industry prosperity index of the to-be-analyzed industry according to the predicted index data;

[0049] determining corresponding predicted power demand information of the to-be-analyzed industry based on the industry prosperity index by using a trained power demand information prediction model.

[0050] In a fourth aspect, the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the following steps:

[0051] in response to a power demand prediction request for a to-be-analyzed industry, acquiring preset industry indexes corresponding to the to-be-analyzed industry;

[0052] performing index screening processing on the preset industry indexes to obtain target industry indexes corresponding to the to-be-analyzed industry;

[0053] obtaining current index data of the to-be-analyzed industry under the target industry index;

[0054] inputting the current index data into the trained index data prediction model to obtain predicted index data of the to-be-analyzed industry under the target industry index;

[0055] determining a corresponding prosperity index of the to-be-analyzed industry according to the predicted index data;

[0056] determining corresponding predicted power demand information of the to-be-analyzed industry based on the prosperity index through the trained power demand information prediction model.

[0057] In a fifth aspect, the present application further provides a computer program product comprising a computer program which, when executed by a processor, implements the following steps:

[0058] obtaining preset industry indexes corresponding to the to-be-analyzed industry in response to a power demand prediction request for the to-be-analyzed industry;

[0059] performing index screening processing on the preset industry indexes to obtain target industry indexes corresponding to the to-be-analyzed industry;

[0060] obtaining current index data of the to-be-analyzed industry under the target industry index;

[0061] inputting the current index data into the trained index data prediction model to obtain predicted index data of the to-be-analyzed industry under the target industry index;

[0062] determining a corresponding prosperity index of the to-be-analyzed industry according to the predicted index data;

[0063] determining corresponding predicted power demand information of the to-be-analyzed industry based on the prosperity index through the trained power demand information prediction model.

[0064] The power demand prediction method, device, computer device, storage medium and computer program product based on the industry prosperity index, first, in response to a power demand prediction request for an industry to be analyzed, a preset industry index corresponding to the industry to be analyzed is obtained, and the preset industry index is subjected to index screening processing to obtain a target industry index corresponding to the industry to be analyzed, then, current index data of the industry to be analyzed under the target industry index is obtained, then, the current index data is input into the trained index data prediction model to obtain predicted index data of the industry to be analyzed under the target industry index, then, according to the predicted index data, a prosperity index corresponding to the industry to be analyzed is determined, and finally, through the trained power demand information prediction model, the predicted power demand information corresponding to the industry to be analyzed is determined based on the prosperity index. In this way, when predicting the power demand, the target industry index that fits the industry to be analyzed is selected from the preset industry index, ensuring the effectiveness of data input, and then the predicted index data predicted by the trained model, so that the prosperity index corresponding to the industry to be analyzed can be more accurately determined, the objective quantitative judgment of the industry operation state is realized, and then the predicted power demand information is accurately matched based on the prosperity index, so that the power demand prediction and the industry prosperity degree are accurately linked, which is beneficial to improve the prediction accuracy of the power demand. Moreover, the whole process combines the influence of economic situation change on power consumption, avoids the defects that the time series analysis, machine learning and the like easily ignore the influence of economic situation change on power consumption, and further improves the prediction accuracy of the power demand. BRIEF DESCRIPTION OF DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the description of the embodiments of the present application or the related art will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other related drawings can be obtained by those skilled in the art without creative labor.

[0066] Figure 1 A flowchart of an industry prosperity index-based power demand prediction method in one embodiment;

[0067] Figure 2 A flowchart of an industry prosperity index-based power demand prediction method in another embodiment;

[0068] Figure 3 A flowchart of an industry prosperity index method based on high-frequency data low-frequency in one embodiment;

[0069] Figure 4 A structural block diagram of an industry prosperity index-based power demand prediction device in one embodiment;

[0070] Figure 5 Fig. 1 is a schematic diagram of the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0071] For the purpose of making the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and should not be used to limit the present application.

[0072] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0073] In one exemplary embodiment, as shown in Figure 1 Fig. 1, a power demand prediction method based on industry index is provided, and the present embodiment takes the method applied to a server as an example; it can be understood that the method can also be applied to a terminal, and can also be applied to a system including a terminal and a server, and is realized through the interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, notebook computers, smart phones and tablet computers; the server can be realized by an independent server or a server cluster composed of multiple servers. In the present embodiment, the method includes the following steps:

[0074] Step S101, in response to a power demand prediction request for an industry to be analyzed, acquiring preset industry indicators corresponding to the industry to be analyzed.

[0075] The industry to be analyzed refers to a specific industry that needs to be analyzed, such as, for example, new energy automobile manufacturing industry, catering industry, steel industry, etc.

[0076] The power demand prediction request refers to a request for power demand prediction of the industry to be analyzed.

[0077] The preset industry indicators refer to high-frequency economic indicators preset for the industry to be analyzed.

[0078] Exemplarily, the server receives a power demand prediction request for an industry to be analyzed sent by a terminal, and performs rationality verification processing on the power demand prediction request to obtain a verification result; in a case where the verification result indicates that the power demand prediction request passes, the server analyzes the power demand prediction request in response to the power demand prediction request, obtains industry category information of the industry to be analyzed, and acquires, from the index mapping rule library, indexes corresponding to the industry category information as preset industry indexes corresponding to the industry to be analyzed.

[0079] In step S102, index screening processing is performed on the preset industry indexes to obtain target industry indexes corresponding to the industry to be analyzed.

[0080] The target industry index refers to a final index obtained after the index screening processing is performed on the preset industry index.

[0081] Exemplarily, the server performs feature extraction processing on the preset industry indexes to obtain feature vectors of the preset industry indexes; then, the feature vectors of the preset industry indexes are input into a plurality of trained importance prediction models to obtain a plurality of predicted importance degrees corresponding to the preset industry indexes; then, the plurality of predicted importance degrees corresponding to the preset industry indexes are summed according to model weights of each trained importance prediction model to obtain a target importance degree corresponding to the preset industry index; then, from the preset industry indexes, a preset industry index with a target importance degree greater than a preset importance degree is selected as a target industry index corresponding to the industry to be analyzed.

[0082] In step S103, current index data of the industry to be analyzed under the target industry index is acquired.

[0083] The current index data is used to represent index data of the industry to be analyzed in a current time under the target industry index.

[0084] Exemplarily, the server determines a data interface corresponding to the target industry index, and acquires the current index data of the industry to be analyzed under the target industry index through the data interface.

[0085] In step S104, the current index data is input into a trained index data prediction model to obtain predicted index data of the industry to be analyzed under the target industry index.

[0086] The index data prediction model refers to a network model capable of obtaining predicted index data of the industry to be analyzed under the target industry index by using the current index data, such as a long short-term memory network model.

[0087] The predicted index data is used to represent index data of the industry to be analyzed in a future time (such as 3 months in the future) under the target industry index.

[0088] Exemplarily, the server denoises the current indicator data to obtain denoised current indicator data; then, the server inputs the denoised current indicator data into the feature extraction model to perform feature extraction processing, to obtain a feature vector of the denoised current indicator data; then, the server inputs the feature vector of the denoised current indicator data into the trained indicator data prediction model to obtain first predicted indicator data of the target industry under the target industry indicator of the industry to be analyzed, and inputs the feature vector of the denoised current indicator data into the historical indicator data prediction model to obtain second predicted indicator data of the target industry under the target industry indicator of the industry to be analyzed; then, the server performs summation processing on the first predicted indicator data and the second predicted indicator data according to the first model weight of the trained indicator data prediction model and the second model weight of the historical indicator data prediction model, to obtain predicted indicator data of the target industry under the target industry indicator of the industry to be analyzed.

[0089] In step S105, the prosperity index corresponding to the industry to be analyzed is determined according to the predicted indicator data.

[0090] The prosperity index refers to a core index reflecting the overall operation state and development trend of the industry in a future period.

[0091] Exemplarily, the server queries the corresponding relationship between the predicted indicator data and the prosperity index based on the predicted indicator data, to obtain the prosperity index corresponding to the predicted indicator data as the prosperity index corresponding to the industry to be analyzed.

[0092] In step S106, the predicted power demand information corresponding to the industry to be analyzed is determined based on the prosperity index by using the trained power demand information prediction model.

[0093] The power demand information prediction model refers to a network model, such as a convolutional neural network model, which can obtain the predicted power demand information corresponding to the industry to be analyzed by using the current prosperity interval.

[0094] The predicted power demand information is used to represent the quantitative result of the power demand of the industry to be analyzed in the future, including predicted demand size (such as monthly predicted power consumption in the next 3 months, annual predicted power consumption size), predicted demand trend (such as capacity increase / decrease / stability, demand growth rate), predicted demand type (such as peak power consumption period distribution, special power consumption scene demand), and other key information.

[0095] Exemplarily, the server determines the predicted demand size, the predicted demand trend, and the predicted demand type corresponding to the industry to be analyzed based on the prosperity index by using the trained power demand information prediction model, which are all used as the predicted power demand information corresponding to the industry to be analyzed.

[0096] In the power demand prediction method based on the industry prosperity index, in response to a power demand prediction request for an industry to be analyzed, preset industry indicators corresponding to the industry to be analyzed are obtained, and the preset industry indicators are subjected to index screening processing to obtain target industry indicators corresponding to the industry to be analyzed. Current index data of the industry to be analyzed under the target industry indicators are obtained. The current index data are input into a trained index data prediction model to obtain predicted index data of the industry to be analyzed under the target industry indicators. Then, the prosperity index corresponding to the industry to be analyzed is determined according to the predicted index data. Finally, the trained power demand information prediction model is used to determine the predicted power demand information corresponding to the industry to be analyzed based on the prosperity index. In this way, when predicting power demand, the target industry indicators that fit the industry to be analyzed are selected from the preset industry indicators to ensure the effectiveness of data input. The predicted index data predicted by the trained model can more accurately determine the prosperity index corresponding to the industry to be analyzed, objectively quantify the judgment of the industry operation state, and accurately match the predicted power demand information based on the prosperity index, so as to link the power demand prediction and the industry prosperity degree accurately, which is beneficial to improving the prediction accuracy of power demand. Moreover, the entire process takes into account the influence of economic situation changes on power consumption, avoiding the defects of time series analysis, machine learning and other methods that easily ignore the influence of economic situation changes on power consumption, and further improving the prediction accuracy of power demand.

[0097] In one exemplary embodiment, the step S102 of performing index screening processing on the preset industry indicators to obtain target industry indicators corresponding to the industry to be analyzed specifically includes the following contents: a first feature vector of the preset industry indicators is extracted, and a second feature vector corresponding to preset industry information of the industry to be analyzed is extracted; the similarity between the preset industry indicators and the preset industry information is determined according to the first feature vector and the second feature vector; from each preset industry indicator, a first candidate industry indicator with a similarity greater than a preset similarity to the preset industry information is selected; the data missing rate and the index coverage rate corresponding to the first candidate industry indicator are obtained, and from each first candidate industry indicator, a second candidate industry indicator with a data missing rate less than a preset data missing rate and an index coverage rate greater than a preset industry index coverage rate is selected; the second candidate industry indicator is subjected to index screening processing to obtain the target industry indicator.

[0098] The first feature vector refers to a feature vector corresponding to the preset industry indicator.

[0099] The preset industry information can refer to information related to the economic logic of the industry to be analyzed, such as core business link information (such as upstream raw material procurement, midstream production and manufacturing, and downstream market demand), key influence factor information (such as raw material price fluctuations and terminal demand changes).

[0100] The second feature vector refers to a feature vector corresponding to the preset industry information.

[0101] The similarity can refer to a cosine similarity.

[0102] The preset similarity refers to a preset similarity threshold.

[0103] The first candidate industry indicator refers to an indicator in the preset industry indicator whose similarity with the preset industry information is greater than the preset similarity.

[0104] The data missing rate refers to a proportion of missing data records in total records of the first candidate industry indicator in a preset statistical period (such as the last 12 months) (for example, if there are 2 months of data missing in a certain indicator in the last 12 months, the data missing rate is 16.7%).

[0105] The indicator coverage rate refers to a proportion of a business range or a sample size of the industry to be analyzed that can be covered by the first candidate industry indicator (for example, if the “photovoltaic installation capacity” indicator covers more than 80% of photovoltaic projects in a preset region, the coverage rate is 80%).

[0106] The preset data missing rate refers to a preset data missing rate threshold (such as 5%).

[0107] The preset industry indicator coverage rate refers to a preset indicator coverage rate threshold (such as 70%).

[0108] The second candidate industry indicator refers to an indicator in the first candidate industry indicator whose data missing rate is less than the preset data missing rate and whose indicator coverage rate is greater than the preset industry indicator coverage rate.

[0109] Exemplarily, the server obtains core business link information and key influence factor information of an industry to be analyzed from a database, both as preset industry information of the industry to be analyzed; then, the preset industry indicators and the preset industry information are input into a feature extraction model respectively, a first feature vector of the preset industry indicators is extracted by the feature extraction model, and a second feature vector corresponding to the preset industry information is extracted; then, the first feature vector and the second feature vector are input into a similarity prediction model to obtain a similarity between the preset industry indicators and the preset industry information; then, from the preset industry indicators, a first candidate industry indicator with a similarity greater than a preset similarity to the preset industry information is screened out; then, all theoretical records of the first candidate industry indicator in a preset statistical period are retrieved from a data source database according to the preset statistical period, records with no data collected, null values or invalid values are screened out, and a proportion of the number of records with missing data to the total number of records is taken as a data missing rate corresponding to the first candidate industry indicator; then, a ratio between a business range actually covered by the first candidate industry indicator and a business range of the industry to be analyzed is taken as an indicator coverage rate corresponding to the first candidate industry indicator; then, from the first candidate industry indicators, a second candidate industry indicator with a data missing rate less than a preset data missing rate and an indicator coverage rate greater than a preset industry indicator coverage rate is screened out; then, the second candidate industry indicator is subjected to indicator screening processing to obtain a target industry indicator.

[0110] In this embodiment, through multi-round progressive screening of feature vector matching, data quality verification and accurate purification, efficient optimization of the preset industry indicators to the target industry indicators is realized, reliable data support is provided for subsequent indicator data prediction, business index calculation and power demand prediction, and problems such as subjective experience judgment, uneven data quality, poor adaptability of indicators to industries in traditional indicator screening are effectively avoided.

[0111] In one exemplary embodiment, the second candidate industry indicators are subjected to indicator screening processing to obtain the target industry indicators, specifically including the following contents: the correlation between the second candidate industry indicators is obtained; the second candidate industry indicators are subjected to indicator screening processing according to the correlation between the second candidate industry indicators to obtain third candidate industry indicators; the correlation between the third candidate industry indicators is less than a preset correlation; the variance inflation factor values between the third candidate industry indicators are determined according to the regression determination coefficients between the third candidate industry indicators; the third candidate industry indicators are subjected to again indicator screening processing according to the variance inflation factor values between the third candidate industry indicators to obtain fourth candidate industry indicators; the variance inflation factor values between the fourth candidate industry indicators are less than a preset variance inflation factor value; and the target industry indicators are obtained based on the fourth candidate industry indicators.

[0112] The correlation is used to represent a quantitative value of a degree of linear correlation between each candidate industry indicator.

[0113] The third candidate industry indicator is an indicator set filtered from the second candidate industry indicator, and the correlation between each two indicators is less than the preset correlation.

[0114] The preset correlation is a preset correlation threshold. It should be noted that the preset correlation is determined according to the situation.

[0115] The regression determination coefficient is a statistical quantity representing an explanatory degree of a dependent variable when a certain candidate industry indicator is used as the dependent variable and a set of independent variables is composed of other candidate industry indicators, and the value range is [0, 1].

[0116] The variance inflation factor value is a quantitative value for measuring a degree of multicollinearity between a single candidate industry indicator and other candidate industry indicators.

[0117] The preset variance inflation factor value is a preset variance inflation factor value threshold. It should be noted that the preset variance inflation factor value is determined according to the situation.

[0118] For example, the server inputs the second candidate industry indicators into the correlation detection model to obtain the correlation between each second candidate industry indicator. Then, from each second candidate industry indicator, a candidate industry indicator combination with a correlation between each second candidate industry indicator greater than or equal to a preset correlation is identified, and the candidate industry indicator with the highest accuracy corresponding to the candidate industry indicator combination is retained to obtain third candidate industry indicators, so that the correlation between each third candidate industry indicator is less than the preset correlation. Then, the third candidate industry indicators are input into the regression determination coefficient determination model to obtain the regression determination coefficient between each third candidate industry indicator. According to the regression determination coefficient between each third candidate industry indicator, the corresponding relationship between the regression determination coefficient and the variance inflation factor value is queried to obtain the variance inflation factor value between each third candidate industry indicator. Then, according to the variance inflation factor value between each third candidate industry indicator, a candidate industry indicator combination with a variance inflation factor value between each third candidate industry indicator greater than or equal to a preset variance inflation factor value is identified, and the candidate industry indicator with the highest accuracy corresponding to the candidate industry indicator combination is retained to obtain fourth candidate industry indicators, so that the variance inflation factor value between each fourth candidate industry indicator is less than the preset variance inflation factor value. Then, the fourth candidate industry indicators are used as target industry indicators.

[0119] In this embodiment, through two rounds of progressive screening of correlation de-redundancy and collinearity interference reduction, the precision purification of candidate industry indicators is realized, which not only retains the core features with high correlation to the industry and up-to-standard data quality, but also completely avoids the problems of information overlap and mutual interference between indicators. A more detailed high-frequency indicator screening and multiple collinearity control mechanism is introduced, which provides more accurate and reliable core data support for subsequent indicator data prediction, calculation of the prosperity index and power demand prediction.

[0120] In one exemplary embodiment, before the step S104 of inputting the current indicator data into the trained indicator data prediction model to obtain the predicted indicator data of the industry to be analyzed under the target industry indicator, the following content is specifically included: smoothing the current indicator data to obtain processed current indicator data; the data frequency of the processed current indicator data is less than the data frequency of the current indicator data.

[0121] Therefore, the step S104 of inputting the current indicator data into the trained indicator data prediction model to obtain the predicted indicator data of the industry to be analyzed under the target industry indicator specifically includes the following content: inputting the processed current indicator data into the trained indicator data prediction model to obtain the predicted indicator data of the industry to be analyzed under the target industry indicator.

[0122] The processed current indicator data refers to the current indicator data after smoothing. It should be noted that the current indicator data belongs to high-frequency economic indicator data, and the processed current indicator data belongs to low-frequency economic indicator data.

[0123] For example, the server extracts the data features of the current indicator data (such as data frequency, fluctuation type, and trend characteristics), and determines the smoothing model corresponding to the current indicator data according to these features. Using this smoothing model, the server smooths the current indicator data to obtain processed current indicator data, ensuring that the data frequency of the processed current indicator data is lower than the current indicator data frequency. Next, the server performs denoising processing on the processed current indicator data to obtain denoised current indicator data. Finally, the server inputs the denoised current indicator data into a feature extraction model for feature extraction processing to obtain denoised current indicator data. The server inputs the feature vectors of the current indicator data after denoising into the trained indicator data prediction model to obtain the first predicted indicator data of the industry to be analyzed under the target industry indicator. It also inputs the feature vectors of the current indicator data after denoising into the historical indicator data prediction model to obtain the second predicted indicator data of the industry to be analyzed under the target industry indicator. Next, the server sums the first and second predicted indicator data according to the first model weights of the trained indicator data prediction model and the second model weights of the historical indicator data prediction model to obtain the predicted indicator data of the industry to be analyzed under the target industry indicator.

[0124] In this embodiment, by smoothing the data frequency, short-term irrelevant interference can be filtered out, the core trend of indicator changes can be highlighted, the data can be made more stable and reliable, and the impact of random fluctuations in high-frequency data can be further weakened, providing more accurate, efficient and high-quality data support that meets actual analysis needs for subsequent economic climate index calculation and electricity demand forecasting.

[0125] In an exemplary embodiment, step S105 above, which determines the prosperity index corresponding to the industry to be analyzed based on the predicted indicator data, specifically includes the following: normalizing the predicted indicator data based on the historical indicator data corresponding to the predicted indicator data to obtain the processed predicted indicator data; determining the scaling factor corresponding to the processed predicted indicator data based on the historical indicator data; and querying a preset correspondence based on the processed predicted indicator data and the scaling factor to obtain the prosperity index corresponding to the industry to be analyzed.

[0126] Historical indicator data refers to the indicator data of the industry to be analyzed within a historical period under the target industry indicators.

[0127] Among them, the processed forecast index data refers to the forecast index data after normalization processing.

[0128] The scaling factor is used to characterize the mapping ratio between historical indicator data and processed predicted indicator data.

[0129] Among them, the preset correspondence is used to represent the correspondence between the processed forecast indicator data, scaling factor and prosperity index.

[0130] For example, the business climate index can be calculated using the following formula (i.e., a pre-defined correspondence):

[0131] Equation (1)

[0132] in, This refers to the business climate index. This refers to the scaling factor. This refers to the processed predictive indicator data.

[0133] For example, the server obtains historical indicator data corresponding to the predicted indicator data, extracts the data characteristics of the historical indicator data (such as data frequency, fluctuation type, and trend characteristics), and determines the normalization processing model corresponding to the predicted indicator data according to the data characteristics of the historical indicator data. For example, if the historical indicator data is relatively uniformly distributed, the Min-Max (minimum value minus maximum value) normalization method is used; if the historical data has extreme values, the Z-Score (z-score) normalization method is selected. Then, the predicted indicator data is input into the normalization processing model for normalization processing to obtain the processed predicted indicator data. Then, based on the key parameters such as the mean, maximum value, minimum value, and standard deviation of the historical indicator data, and based on these key parameters, the scaling factor corresponding to the processed predicted indicator data is determined. Then, based on the processed predicted indicator data and the scaling factor, a preset correspondence is queried to obtain the prosperity index corresponding to the industry to be analyzed. The preset correspondence is used to represent the correspondence between the processed predicted indicator data, the scaling factor, and the prosperity index.

[0134] In this embodiment, the dimensional differences of the predicted indicator data are eliminated by normalization processing driven by historical indicator data. The correlation between the data and the actual operation of the industry is calibrated by the corresponding scaling factor. Then, the prosperity index is obtained by accurate mapping through the preset correspondence relationship. This realizes the efficient transformation of the predicted data into the industry prosperity status, making the prosperity index more intuitive and interpretable, and effectively avoiding the errors caused by subjective experience judgment.

[0135] In an exemplary embodiment, step S106 above, using a trained electricity demand information prediction model and based on a business climate index, determines the predicted electricity demand information corresponding to the industry to be analyzed. Specifically, this includes: determining the current business climate interval corresponding to the business climate index; inputting the current business climate interval into the trained electricity demand information prediction model to obtain the predicted probability of the industry to be analyzed under each preset electricity demand information; and selecting the preset electricity demand information with the highest predicted probability from each preset electricity demand information as the predicted electricity demand information corresponding to the industry to be analyzed.

[0136] Among them, the current economic climate range refers to the specific range to which the economic climate index belongs, including the overheated range, the slightly overheated range, the stable range, the slightly underheated range, and the underheated range.

[0137] Among them, the preset electricity demand information refers to the pre-set electricity demand information, including the preset demand scale, preset demand trend and preset demand type.

[0138] Among them, the prediction probability refers to the likelihood that the electricity demand information prediction model determines the preset electricity demand information to be correct.

[0139] For example, the server determines the correspondence between the prosperity index and the current prosperity interval based on the scaling factor corresponding to the processed predicted index data, and queries the correspondence based on the prosperity index to obtain the current prosperity interval corresponding to the prosperity index; then, it extracts the upper limit and lower limit of the current prosperity interval, and performs feature extraction processing on the upper limit and lower limit respectively to obtain the feature vectors of the upper limit and lower limit; then, it fuses the feature vectors of the upper limit and lower limit to obtain a fused feature vector, and inputs the fused feature vector into the trained electricity demand information prediction model to obtain the industry to be analyzed in each preset demand range. The system first calculates the predicted probability under a given model, the predicted probability of the industry under each preset demand trend, and the predicted probability of the industry under each preset demand type. Then, it selects the preset demand scale with the highest predicted probability from each preset demand scale as the predicted demand scale corresponding to the industry under analysis; the preset demand trend with the highest predicted probability from each preset demand trend as the predicted demand trend corresponding to the industry under analysis; and the preset demand type with the highest predicted probability from each preset demand type as the predicted demand type corresponding to the industry under analysis. Finally, the predicted demand scale, predicted demand trend, and predicted demand type corresponding to the industry under analysis are all used as the predicted electricity demand information for the industry under analysis.

[0140] In this embodiment, by introducing the division of the prosperity index interval and adopting the symmetrical interval division method based on standard deviation, the degree of industry expansion and contraction is more intuitively reflected. Furthermore, by using the trained power demand information prediction model, the prediction probability of each preset power demand information is accurately output. Finally, the preset power demand information with the highest probability is selected as the result, which effectively avoids the misjudgment of demand caused by subjective judgment or single-dimensional analysis, and provides a reliable basis for the accurate generation of subsequent power demand prediction instructions.

[0141] In one exemplary embodiment, such as Figure 2 As shown, another method for forecasting electricity demand based on an industry prosperity index is provided. Taking the application of this method to a server as an example, the specific steps include:

[0142] Step S201: In response to the power demand forecast request for the industry to be analyzed, obtain the preset industry indicators corresponding to the industry to be analyzed.

[0143] Step S202: Extract the first feature vector of the preset industry indicator and the second feature vector corresponding to the preset industry information of the industry to be analyzed; determine the similarity between the preset industry indicator and the preset industry information based on the first feature vector and the second feature vector.

[0144] Step S203: From each preset industry indicator, select the first candidate industry indicator whose similarity to the preset industry information is greater than the preset similarity; obtain the data missing rate and indicator coverage corresponding to the first candidate industry indicator, and from each first candidate industry indicator, select the second candidate industry indicator whose data missing rate is less than the preset data missing rate and whose indicator coverage is greater than the preset industry indicator coverage.

[0145] Step S204: Obtain the correlation between each second candidate industry indicator; based on the correlation between each second candidate industry indicator, perform indicator screening processing on the second candidate industry indicators to obtain the third candidate industry indicators; the correlation between each third candidate industry indicator is less than the preset correlation.

[0146] Step S205: Determine the variance inflation factor value among the third candidate industry indicators based on the regression determination coefficients among the third candidate industry indicators.

[0147] Step S206: Based on the variance inflation factor values ​​between each third candidate industry indicator, the third candidate industry indicators are further screened to obtain fourth candidate industry indicators; the variance inflation factor values ​​between each fourth candidate industry indicator are all less than the preset variance inflation factor value; based on the fourth candidate industry indicators, the target industry indicators corresponding to the industry to be analyzed are obtained.

[0148] Step S207: Obtain the current indicator data of the industry to be analyzed under the target industry indicators.

[0149] Step S208: Smooth the current indicator data to obtain the processed current indicator data; the data frequency of the processed current indicator data is less than the data frequency of the current indicator data.

[0150] Step S209: Input the processed current indicator data into the trained indicator data prediction model to obtain the predicted indicator data of the industry to be analyzed under the target industry indicator.

[0151] Step S210: Normalize the predicted indicator data according to the historical indicator data corresponding to the predicted indicator data to obtain the processed predicted indicator data; determine the scaling factor corresponding to the processed predicted indicator data according to the historical indicator data.

[0152] Step S211: Based on the processed forecast indicator data and scaling factor, query the preset correspondence to obtain the prosperity index corresponding to the industry to be analyzed; the preset correspondence is used to represent the correspondence between the processed forecast indicator data, scaling factor and prosperity index.

[0153] Step S212: Determine the current economic climate range corresponding to the economic climate index; input the current economic climate range into the trained electricity demand information prediction model to obtain the predicted probability of the industry to be analyzed under each preset electricity demand information.

[0154] Step S213: Select the preset power demand information with the highest prediction probability from the preset power demand information, and use it as the predicted power demand information for the industry to be analyzed.

[0155] In the aforementioned electricity demand forecasting method based on industry prosperity indices, target industry indicators that fit the industry under analysis are selected from preset industry indicators to ensure the validity of data input. The forecast indicator data is then predicted using a trained model, which more accurately determines the prosperity index corresponding to the industry under analysis. This enables an objective quantitative assessment of the industry's operating status, and the electricity demand information is accurately matched and predicted based on the prosperity index, allowing for precise linkage between electricity demand forecasting and industry prosperity, thus improving the accuracy of electricity demand forecasting. Furthermore, the entire process incorporates the impact of economic changes on electricity consumption, avoiding the shortcomings of time series analysis and machine learning methods that easily overlook the impact of economic changes on electricity consumption and thus affect the accuracy of electricity demand forecasting. This further improves the accuracy of electricity demand forecasting.

[0156] In an exemplary embodiment, to more clearly illustrate the electricity demand forecasting method based on the industry prosperity index provided in this application, the following specific embodiment will be used to describe the electricity demand forecasting method based on the industry prosperity index. In one embodiment, as... Figure 3 As shown, this application also provides a method for constructing an industry prosperity index based on the low-frequency conversion of high-frequency data. Specifically, it includes the following:

[0157] 1. High-frequency indicator screening:

[0158] Based on data availability, economic logic, and industry representativeness, indicators were initially selected from multiple dimensions. To ensure that high-frequency indicators can effectively explain the fluctuations of the target variable and to avoid redundancy and excessive noise in the model, further screening of candidate high-frequency indicators is required, specifically including the following two steps:

[0159] (1) Correlation detection; First, the correlation between candidate high-frequency indicators is tested. Based on the Pearson correlation coefficient, that is, for two variables... Define the correlation coefficient:

[0160] Equation (2)

[0161] in, For variables covariance, For variables The standard deviation of the correlation coefficient. The correlation coefficient ranges from [-1, 1]. The correlation matrix can clearly show the pairwise correlations between all variables at once.

[0162] Equation (3)

[0163] If the absolute value is higher than the set threshold (e.g., 0.8), it indicates that the two indicators have highly overlapping information. In this case, the more representative or stable indicator is retained, and the redundant indicator is identified.

[0164] (2) Multicollinearity test; secondly, to avoid linear dependency relationships among multiple indicators, this method further introduces the variance inflation factor (VIF) for testing:

[0165] Equation (4)

[0166] in, Let be the regression determination coefficient when the i-th variable is the dependent variable and the other variables are independent variables. If If the value is greater than 10, it indicates that the indicator has strong multicollinearity with other variables and should be removed.

[0167] 2. High-frequency to low-frequency conversion:

[0168] After selecting high-frequency indicators, they need to be converted into low-frequency indicators (such as monthly indicators) to align with the low-frequency target variable. Smoothing is then performed using a moving average method.

[0169] Equation (5)

[0170] in, The converted low-frequency indicator represents the high-frequency observations within the window, and k is the window length (e.g., the number of weeks or days in a month). The resulting low-frequency index maintains the same time frequency as the target variable, allowing it to be directly used for subsequent regression modeling.

[0171] 3. Rolling window regression estimation:

[0172] After performing low-frequency reduction, a rolling regression approach is used to capture dynamic relationships. First, a fixed time window is set (e.g., x years, T months). Within each window, the low-frequency target variable... As the dependent variable, the processed high-frequency index sequence (i represents different indicators) are used as independent variables for multiple regression estimation:

[0173] Equation (6)

[0174] in, For the estimated regression coefficients, This represents the error term. Regression modeling is performed by substituting historical data within the window length, and then the obtained coefficients are used to apply to the out-of-sample data for the next period. Prediction is made to obtain a continuous sequence of predicted values.

[0175] 4. Exponentialization of the predicted value sequence:

[0176] To ensure the comparability of prediction results across different indicators and time periods, the predicted value series needs to be standardized. Here, the Z-score standardization method is used:

[0177] Equation (7)

[0178] Where μ is the mean of the baseline period (i.e., historical data), The standard deviation for the base period represents the range of fluctuation. Standardization eliminates the influence of dimensions and units, transforming values ​​into relative positions, i.e., "how many standard deviations above the mean" or "how many standard deviations below the mean." This approach ensures that comparisons can be made on the same scale regardless of the magnitude of the original indicator values.

[0179] After obtaining the standardized Z-score series, we further convert it into a business climate index with 50 as the central value, i.e., using the formula:

[0180] Equation (1)

[0181] Among them, the central value is an internationally accepted practice (such as the PMI index), which can intuitively distinguish the state of the economy: above 50 indicates that the industry is in the expansion range, and below 50 indicates that it is in the contraction range. Parameters This is an optional scaling factor that determines the fluctuation range of the business climate index during the base period, which is equivalent to mapping the standard deviation of (6) to the number of scale points of the business climate index; it is usually set to 1 by default, that is, the standard deviation of the business climate index during the base period is 1. In this formula, This represents the multiple of the standard deviation of the current value relative to the mean of the base period, while This corresponds to the fluctuation range under the business climate index scale. Its function is to define the statistical definition The standard deviation is mapped to s points on the business climate index scale.

[0182] 5. Division of the economic climate index range:

[0183] After indexation, the economic climate index needs to be divided into intervals to distinguish different economic states. A symmetrical interval division method based on standard deviation is adopted, that is, with 50 as the center, the intervals are divided symmetrically upwards and downwards according to multiples of the standard deviation.

[0184] The final output is a continuous series of predicted business climate indices, along with economic status classifications for different intervals. It can be used for real-time monitoring, policy evaluation, and industry analysis.

[0185] By differentiating industries with varying levels of prosperity and conducting business expansion analysis, we can accurately grasp the growth points of electricity sales, extract and predict the capacity increase and decrease of various industries, realize full-process monitoring of customer business expansion, and promote the improvement and acceleration of business expansion and power supply.

[0186] In the above embodiments, when forecasting electricity demand, target industry indicators that fit the industry to be analyzed are selected from preset industry indicators to ensure the validity of data input. Then, the predicted indicator data is obtained through a trained model, thereby more accurately determining the corresponding prosperity index of the industry to be analyzed. This enables an objective quantitative assessment of the industry's operating status, and based on the prosperity index, electricity demand information is accurately matched and predicted, allowing for precise linkage between electricity demand forecasting and industry prosperity, which helps improve the accuracy of electricity demand forecasting. Furthermore, the entire process incorporates the impact of economic changes on electricity consumption, avoiding the shortcomings of time series analysis and machine learning, which tend to overlook the impact of economic changes on electricity consumption and thus affect the accuracy of electricity demand forecasting. This further improves the accuracy of electricity demand forecasting. Simultaneously, by utilizing high-frequency data and processing it with low-frequency data, the shortcomings of insufficient low-frequency forecast samples and strong lag are compensated for. Regression modeling and indexation improve the stability and interpretability of the forecast. Different target variables are set for macroeconomic monitoring and analysis of different industries.

[0187] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0188] Based on the same inventive concept, this application also provides an industry-based electricity demand forecasting device for implementing the above-mentioned industry-based electricity demand forecasting method. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more industry-based electricity demand forecasting device embodiments provided below can be found in the limitations of the industry-based electricity demand forecasting method described above, and will not be repeated here.

[0189] In one exemplary embodiment, such as Figure 4 As shown, a power demand forecasting device based on an industry prosperity index is provided, comprising: an index acquisition module 401, an index screening module 402, a data acquisition module 403, a data forecasting module 404, an index determination module 405, and a demand forecasting module 406, wherein:

[0190] The indicator acquisition module 401 is used to obtain the preset industry indicators corresponding to the industry to be analyzed in response to the power demand forecast request for the industry to be analyzed.

[0191] The indicator filtering module 402 is used to filter preset industry indicators to obtain the target industry indicators corresponding to the industry to be analyzed.

[0192] The data acquisition module 403 is used to acquire the current indicator data of the industry to be analyzed under the target industry indicators.

[0193] The data prediction module 404 is used to input the current indicator data into the trained indicator data prediction model to obtain the predicted indicator data of the industry to be analyzed under the target industry indicator.

[0194] The index determination module 405 is used to determine the prosperity index corresponding to the industry to be analyzed based on the forecast indicator data.

[0195] The demand forecasting module 406 is used to determine the predicted electricity demand information for the industry to be analyzed based on the business climate index, using the trained electricity demand information forecasting model.

[0196] In an exemplary embodiment, the indicator screening module 402 is further configured to extract a first feature vector of a preset industry indicator and a second feature vector corresponding to preset industry information of the industry to be analyzed; determine the similarity between the preset industry indicator and the preset industry information based on the first feature vector and the second feature vector; screen out first candidate industry indicators from each preset industry indicator whose similarity to the preset industry information is greater than a preset similarity; obtain the data missing rate and indicator coverage corresponding to the first candidate industry indicator, and screen out second candidate industry indicators from each first candidate industry indicator whose data missing rate is less than a preset data missing rate and whose indicator coverage is greater than the preset industry indicator coverage; perform indicator screening processing on the second candidate industry indicator to obtain the target industry indicator.

[0197] In an exemplary embodiment, the indicator screening module 402 is further configured to: obtain the correlation between each second candidate industry indicator; perform indicator screening processing on the second candidate industry indicators based on the correlation between each second candidate industry indicator to obtain a third candidate industry indicator; the correlation between each third candidate industry indicator is less than a preset correlation; determine the variance inflation factor value between each third candidate industry indicator based on the regression determination coefficient between each third candidate industry indicator; perform indicator screening processing on the third candidate industry indicators again based on the variance inflation factor value between each third candidate industry indicator to obtain a fourth candidate industry indicator; the variance inflation factor value between each fourth candidate industry indicator is less than a preset variance inflation factor value; and obtain the target industry indicator based on the fourth candidate industry indicator.

[0198] In an exemplary embodiment, the power demand forecasting device based on the industry prosperity index further includes a data processing module for smoothing the current indicator data to obtain processed current indicator data; the data frequency of the processed current indicator data is less than the data frequency of the current indicator data; the data prediction module 404 is further used to input the processed current indicator data into the trained indicator data prediction model to obtain the predicted indicator data of the industry to be analyzed under the target industry indicator.

[0199] In an exemplary embodiment, the index determination module 405 is further configured to normalize the predicted index data according to the historical index data corresponding to the predicted index data to obtain the processed predicted index data; determine the scaling factor corresponding to the processed predicted index data according to the historical index data; and query a preset correspondence based on the processed predicted index data and the scaling factor to obtain the prosperity index corresponding to the industry to be analyzed; the preset correspondence is used to represent the correspondence between the processed predicted index data, the scaling factor and the prosperity index.

[0200] In an exemplary embodiment, the demand forecasting module 406 is further configured to determine the current economic climate interval corresponding to the economic climate index; input the current economic climate interval into the trained electricity demand information forecasting model to obtain the predicted probability of the industry to be analyzed under each preset electricity demand information; and select the preset electricity demand information with the highest predicted probability from each preset electricity demand information as the predicted electricity demand information corresponding to the industry to be analyzed.

[0201] The modules in the aforementioned power demand forecasting device based on industry prosperity indices can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0202] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores current and forecast data. The I / O interfaces allow the processor to exchange information with external devices. The communication interface allows communication with external terminals via a network connection. When executed by the processor, the computer program implements a method for forecasting electricity demand based on an industry prosperity index.

[0203] Those skilled in the art will understand that Figure 5The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0204] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0205] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above-described method embodiments.

[0206] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0207] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0208] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0209] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for forecasting electricity demand based on an industry prosperity index, characterized in that, The method includes: In response to a power demand forecast request for the industry to be analyzed, the preset industry indicators corresponding to the industry to be analyzed are obtained. The preset industry indicators are filtered to obtain the target industry indicators corresponding to the industry to be analyzed. Obtain the current indicator data of the industry to be analyzed under the target industry indicator; The current indicator data is input into the trained indicator data prediction model to obtain the predicted indicator data of the industry to be analyzed under the target industry indicator. Based on the predicted indicator data, the prosperity index corresponding to the industry to be analyzed is determined; Based on the prosperity index, the predicted electricity demand information for the industry to be analyzed is determined by the trained electricity demand prediction model.

2. The method according to claim 1, characterized in that, The step of performing index filtering on the preset industry indicators to obtain the target industry indicators corresponding to the industry to be analyzed includes: Extract the first feature vector of the preset industry indicator, and obtain the second feature vector corresponding to the preset industry information of the industry to be analyzed; Based on the first feature vector and the second feature vector, the similarity between the preset industry index and the preset industry information is determined; From the preset industry indicators, a first candidate industry indicator with a similarity greater than a preset similarity to the preset industry information is selected; Obtain the data missing rate and indicator coverage corresponding to the first candidate industry indicator, and select a second candidate industry indicator from each of the first candidate industry indicators whose data missing rate is less than a preset data missing rate and whose indicator coverage is greater than a preset industry indicator coverage. The second candidate industry indicators are subjected to indicator screening to obtain the target industry indicators.

3. The method according to claim 2, characterized in that, The step of filtering the second candidate industry indicators to obtain the target industry indicators includes: Obtain the correlation between the second candidate industry indicators; Based on the correlation between the second candidate industry indicators, the second candidate industry indicators are screened to obtain third candidate industry indicators; the correlation between the third candidate industry indicators is less than the preset correlation. Based on the regression determination coefficients between the third candidate industry indicators, the variance inflation factor values ​​between the third candidate industry indicators are determined. Based on the variance inflation factor values ​​among the third candidate industry indicators, the third candidate industry indicators are further screened to obtain fourth candidate industry indicators; the variance inflation factor values ​​among the fourth candidate industry indicators are all less than the preset variance inflation factor value. Based on the fourth candidate industry indicator, the target industry indicator is obtained.

4. The method according to claim 1, characterized in that, Before inputting the current indicator data into the trained indicator data prediction model to obtain the predicted indicator data for the industry to be analyzed under the target industry indicator, the process further includes: The current indicator data is smoothed to obtain processed current indicator data; the data frequency of the processed current indicator data is lower than the data frequency of the current indicator data. The step of inputting the current indicator data into the trained indicator data prediction model to obtain the predicted indicator data of the industry to be analyzed under the target industry indicator includes: The processed current indicator data is input into the trained indicator data prediction model to obtain the predicted indicator data of the industry to be analyzed under the target industry indicator.

5. The method according to claim 1, characterized in that, The step of determining the business climate index corresponding to the industry to be analyzed based on the predicted indicator data includes: Based on the historical index data corresponding to the predicted index data, the predicted index data is normalized to obtain the processed predicted index data. Based on the historical indicator data, the scaling factor corresponding to the processed predicted indicator data is determined; Based on the processed forecast indicator data and the scaling factor, a preset correspondence is queried to obtain the business climate index corresponding to the industry to be analyzed; the preset correspondence is used to represent the correspondence between the processed forecast indicator data, the scaling factor and the business climate index.

6. The method according to any one of claims 1 to 5, characterized in that, The electricity demand prediction model, trained through training, determines the predicted electricity demand information for the industry to be analyzed based on the business climate index, including: Determine the current economic climate range corresponding to the aforementioned economic climate index; The current weather forecast interval is input into the trained electricity demand information prediction model to obtain the predicted probability of the industry to be analyzed under each preset electricity demand information. From the preset electricity demand information, the preset electricity demand information with the highest prediction probability is selected as the predicted electricity demand information corresponding to the industry to be analyzed.

7. A power demand forecasting device based on an industry prosperity index, characterized in that, The device includes: The indicator acquisition module is used to acquire preset industry indicators corresponding to the industry to be analyzed in response to a power demand forecast request for the industry to be analyzed. The indicator filtering module is used to perform indicator filtering processing on the preset industry indicators to obtain the target industry indicators corresponding to the industry to be analyzed. The data acquisition module is used to acquire the current indicator data of the industry to be analyzed under the target industry indicator; The data prediction module is used to input the current indicator data into the trained indicator data prediction model to obtain the predicted indicator data of the industry to be analyzed under the target industry indicator. The index determination module is used to determine the prosperity index corresponding to the industry to be analyzed based on the predicted indicator data. The demand forecasting module is used to determine the predicted electricity demand information for the industry to be analyzed based on the prosperity index, using a trained electricity demand information forecasting model.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images