Clustering-based industry subdivision electricity sale quantity prediction method and device, and storage medium

Abstract

The invention relates to a clustering-based industry subdivision electricity sale quantity prediction method and device and a storage medium, which are applied to the technical field of electricity market transactions, and comprise the steps of obtaining representation vectors in a historical preset period of time of an electricity user, so that the electricity consumption behavior of the user can be represented more accurately, the accuracy of a prediction model is improved, and the user experience is improved. Meanwhile, vectorization representation is carried out on the attributes of the users, and the attributes are spliced with power consumption behavior representation vectors, so that more comprehensive user features can be provided through the integration, and the generalization ability of the model is enhanced; according to the method, high-dimensional original time sequence data is converted into low-dimensional feature vectors, so that the calculation burden of a clustering algorithm is reduced, the clustering efficiency is also improved, users are divided into different class clusters, modeling prediction is carried out for the electrical load of each group of users, clearer model interpretation can be provided, and the user experience is improved. And the electricity utilization characteristics of different user groups can be understood conveniently.

Description

Technical Field

[0001] The present invention relates to the technical field of power market transactions, and in particular to a method, device and storage medium for predicting power sales volume by industry segmentation based on clustering. Background Art

[0002] With the opening and development of the domestic electricity market, it is becoming increasingly common for many power sales companies to purchase electricity on behalf of users. For these companies, accurate prediction of electricity consumption and load is the key to ensuring profitability in the electricity market. However, power sales load forecasting faces a series of challenges: the volatility of user group composition, the differences in customer groups of different power sales companies, and the difficulty of information modeling when directly conducting overall power sales load forecasting.

[0003] Existing technologies usually rely on industry classification-based labels to identify users and perform categorized predictions. This method has obvious disadvantages: relying solely on industry information as classification labels is too simplistic and cannot capture the similarities in electricity consumption patterns between different industries, nor can it adapt to the differences in electricity consumption patterns within the same industry, let alone reflect the changes in electricity consumption patterns of the same user in different periods. Static industry label classification methods cannot effectively utilize the similarities in electricity consumption patterns, or incorrectly apply the differences in electricity consumption patterns, ultimately leading to poor modeling results and reduced prediction accuracy. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a clustering-based industry-segmented electricity sales forecasting method, device and storage medium to solve the problem in the prior art that relying solely on industry information as classification labels is too simplistic, cannot capture the similarities in electricity consumption patterns between different industries, cannot adapt to the differences in electricity consumption patterns within the same industry, and cannot reflect the changes in electricity consumption patterns of the same user in different periods. The static industry label classification method cannot effectively utilize the similarities in electricity consumption patterns, or incorrectly applies the differences in electricity consumption patterns, ultimately leading to poor modeling effects and low prediction accuracy.

[0005] According to a first aspect of an embodiment of the present invention, a clustering-based industry segmentation electricity sales forecasting method is provided, the method comprising:

[0006] Acquire historical electricity consumption data of each electricity user; acquire electricity consumption data of each electricity user in a historical preset time period based on the historical electricity consumption data;

[0007] Based on the electricity consumption data of each electricity user in the historical preset time period, obtaining the electricity consumption behavior characterization vector of each electricity user in the historical preset time period;

[0008] Acquire the attributes of each power user, and vectorize the attributes of each power user to obtain the attribute vector of each power user;

[0009] The power consumption behavior characterization vector of each power user and the corresponding self-attribute vector are concatenated to obtain the characterization vector of each power user;

[0010] Based on the characterization vectors of the power users, the power users are clustered using a k-means clustering method to obtain a cluster identifier corresponding to each power user;

[0011] The power users are grouped according to their corresponding cluster identifiers, and the power loads of the power users in the same group are summed up. The pre-built model framework is iteratively trained with the total load obtained by summing up each group of power users to obtain the prediction model corresponding to each group of power users. The prediction model corresponding to each group of power users is used to predict the power load of each power user in the corresponding group.

[0012] Preferably,

[0013] The step of obtaining the power consumption behavior characterization vector of each power user in the historical preset period based on the power consumption data of each power user in the historical preset period includes:

[0014] Setting a window size of the window function, and rolling acquiring a window representation vector in the power consumption data of each power user in a preset historical period based on the preset window function;

[0015] Obtaining a peak-valley value time period vector based on the power consumption data of each power user in a historical preset time period;

[0016] The window characterization vectors and the peak-valley value time period vectors of each power user are spliced ​​to obtain the power consumption behavior characterization vectors of each power user in the historical preset time period.

[0017] Preferably,

[0018] The step of rolling and acquiring the window characterization vector in the power consumption data of each power user in a preset historical period based on a preset window function includes:

[0019] Based on the preset window function, rolling obtain the average power consumption in each window in the historical preset time period, and use the average power consumption as the average value representation vector;

[0020] Based on the preset window function, the maximum power consumption in each window in the historical preset time period is obtained in a rolling manner, and the maximum power consumption is used as a maximum value characterization vector;

[0021] Based on the preset window function, the minimum power consumption in each window in the historical preset time period is obtained in a rolling manner, and the minimum power consumption is used as a minimum value characterization vector;

[0022] The average value characterization vector, the maximum value characterization vector, and the minimum value characterization vector are concatenated to obtain a window characterization vector in the electricity consumption data of each power user in a historical preset time period.

[0023] Preferably,

[0024] The step of obtaining the peak-valley value time period vector based on the power consumption data of each power user in the historical preset time period includes:

[0025] Divide a day into n time periods in units of x hours, obtain the peak power consumption and valley power consumption of each power user in the historical preset time period, and obtain the number of times the peak power consumption and valley power consumption of each power user occur in the corresponding time period, and obtain the peak time period vector and valley time period vector respectively;

[0026] The peak time period vector and the valley time period vector of each power user are spliced ​​together to obtain the peak and valley time period vector of the power consumption data of each power user in the historical preset time period.

[0027] Preferably,

[0028] The attributes of each power user are the industry classification of each power user. The industry classification of each power user is vectorized to obtain the attribute vector of each power user.

[0029] Preferably,

[0030] The electricity consumption data of the historical preset time period is the electricity consumption data within Y days from the day to be predicted.

[0031] According to a second aspect of an embodiment of the present invention, a clustering-based industry segmentation electricity sales forecasting device is provided, the device comprising:

[0032] Historical data acquisition module: used to acquire historical power consumption data of each power user; based on the historical power consumption data, acquire the power consumption data of each power user in a historical preset time period;

[0033] An electricity consumption behavior characterization vector acquisition module: used to acquire the electricity consumption behavior characterization vector of each electricity user in a historical preset period based on the electricity consumption data of each electricity user in a historical preset period;

[0034] Attribute vector acquisition module: used to acquire the attributes of each power user, vectorize the attributes of each power user, and obtain the attribute vector of each power user;

[0035] A characterization vector acquisition module is used to concatenate the characterization vector of the power consumption behavior of each power user and the corresponding self-attribute vector to obtain the characterization vector of each power user;

[0036] Clustering module: used to cluster each power user based on the characterization vector of each power user using the k-means clustering method to obtain a cluster identifier corresponding to each power user;

[0037] Group prediction module: It is used to group the power users according to the cluster identifier corresponding to each power user, sum the power load of the power users in the same group, iteratively train the pre-built model framework with the total load obtained by summing up each group of power users, and obtain the prediction model corresponding to each group of power users. The power load of each power user in the corresponding group is predicted by the prediction model corresponding to each group of power users.

[0038] According to a third aspect of an embodiment of the present invention, there is provided a storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a main controller, each step in the above method is implemented.

[0039] The technical solution provided by the embodiments of the present invention may have the following beneficial effects:

[0040] By obtaining the characterization vectors of the power users in the historical preset time periods, the present application can more accurately characterize the users' electricity consumption behavior, thereby improving the accuracy of the prediction model. At the same time, the user's own attributes are vectorized and spliced ​​with the electricity consumption behavior characterization vector. Such integration can provide more comprehensive user characteristics and enhance the generalization ability of the model; the high-dimensional original time series data is converted into low-dimensional feature vectors, which not only reduces the computational burden of the clustering algorithm, but also improves the efficiency of clustering. By dividing the users into different clusters and modeling and predicting the electricity load of each group of users, a clearer model explanation can be provided, which facilitates the understanding of the electricity consumption characteristics of different user groups; the above-mentioned scheme can more accurately capture the subtle differences in electricity consumption patterns, thereby enhancing the modeling ability and improving the accuracy of electricity sales load forecasting. Through this dynamic and adaptive clustering and modeling method, the present application can not only enhance the market competitiveness of power sales companies, but also provide strong technical support for the stable operation of the power market.

[0041] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0043] Figure 1 is a flow chart of a method for predicting power sales volume by industry segmentation based on clustering according to an exemplary embodiment;

[0044] Figure 2 is a system schematic diagram of an industry-specific electricity sales forecasting device based on clustering according to another exemplary embodiment;

[0045] In the attached figure: 1-historical data acquisition module, 2-electricity consumption behavior characterization vector acquisition module, 3-attribute vector acquisition module, 4-characterization vector acquisition module, 5-clustering module, 6-group prediction module. DETAILED DESCRIPTION

[0046] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0047] Embodiment 1

[0048] Figure 1 is a flow chart of a method for predicting power sales volume by industry segmentation based on clustering according to an exemplary embodiment. Figure 1 As shown, the method includes:

[0049] S1, obtaining historical electricity consumption data of each electricity user; obtaining electricity consumption data of each electricity user in a historical preset time period based on the historical electricity consumption data;

[0050] S2, obtaining a representation vector of the electricity consumption behavior of each electricity user in the historical preset period based on the electricity consumption data of each electricity user in the historical preset period;

[0051] S3, obtaining the attributes of each power user, and vectorizing the attributes of each power user to obtain the attribute vector of each power user;

[0052] S4, concatenating the power consumption behavior representation vector of each power user and the corresponding self-attribute vector to obtain the representation vector of each power user;

[0053] S5, clustering each power user using a k-means clustering method based on the characterization vector of each power user to obtain a cluster identifier corresponding to each power user;

[0054] S6, grouping each power user according to the cluster identifier corresponding to each power user, summing the power loads of the power users in the same group, iteratively training the pre-built model framework with the total load obtained by summing each group of power users, obtaining the prediction model corresponding to each group of power users, and predicting the power load of each power user in the corresponding group with the prediction model corresponding to each group of power users;

[0055] It can be understood that the historical electricity consumption data of proxy users is collected, and the historical electricity consumption period of users that need to be referenced is selected; based on the rolling statistics of the window function and the peak and valley value information, the historical electricity consumption behavior of users is automatically extracted to obtain the electricity consumption behavior representation vector, which specifically includes:

[0056] Window representation vector: Use a sliding window to obtain the average, maximum, and minimum feature extraction corresponding to the user in each window. The window size is selected as the size of a day. The calculation method is as follows:

[0057] Window average: The daily average value of each user is used as the average representation vector:

[0058]

[0059] Window maximum value: Take the daily maximum value of each user as the maximum value representation vector:

[0060]

[0061] Window minimum: The daily minimum value of each user is used as the minimum value representation vector:

[0062]

[0063] In the formula, i represents the i-th day within the date selected as the characteristic period of electricity consumption behavior, i = 1, 2, ..., d, where d is a configurable parameter. In order to ensure the timeliness of the electricity consumption behavior characterization, data within 60 days from the forecast date (i.e., d ≤ 60) is usually selected as the benchmark interval for electricity consumption behavior characterization;

[0064] Peak-valley time period vector: Divide a 24-hour day into six time periods in units of 4 hours, count the peak power consumption and valley power consumption within the benchmark interval, and count the number of times each user's peak power consumption and valley power consumption in the benchmark interval occurs in the corresponding time period;

[0065] As shown below, the peak-valley information of a user can be represented by the peak time period vector and the valley time period vector:

[0066] The peak time vector of a certain user:

[0067]

[0068] The valley time period vector of a certain user:

[0069]

[0070] This embodiment can dynamically adapt to the changes in the user's electricity consumption behavior by extracting rolling windows and peak and valley value information, especially within the benchmark interval for characterizing electricity consumption behavior, selecting data within 60 days from the forecast date as the benchmark, thereby ensuring the timeliness of the characterization of electricity consumption behavior;

[0071] The attributes of the power users are vectorized and represented. In this embodiment, industry classification is selected as the attributes of the power users.

[0072] Methods for vectorizing industry classifications include one-hot encoding, ordered encoding, and binary encoding. Taking one-hot encoding as an example, one-hot encoding refers to converting each industry classification into a binary vector. The length of the vector is equal to the number of classifications. Only one element in the vector is 1, indicating that the sample belongs to this classification, and the other elements are 0. This method is suitable for situations where there is no sequential relationship between classifications. For example, for the classification feature of "manufacturing", "manufacturing" can be encoded as [1,0,0], other industries such as "agriculture" can be encoded as [0,1,0], and "industry" can be encoded as [0,0,1];

[0073] The power consumption behavior representation vector and the user's own attribute vector are concatenated to form a user representation vector;

[0074] Based on the user characterization vector, the k-means clustering method is used to cluster the users to obtain the cluster identifier corresponding to each user; the features used for clustering in this embodiment are based on the statistical features of multiple days, which can to a certain extent eliminate the situation where the temporary power demand of a certain user on a certain day causes the power consumption pattern of the user to be incorrectly clustered;

[0075] It is worth emphasizing that: in the field of data mining and machine learning, time series clustering is an important technique, which involves grouping a series of data points arranged in chronological order in order to discover patterns and trends in the data. However, existing time series clustering methods, such as using the k-means algorithm directly based on the original time series data, have shortcomings in efficiency and effectiveness; for example, the k-means algorithm is computationally inefficient when processing large-scale data sets because it needs to calculate the distance between each data point and all cluster centers, which will significantly increase the computational cost on large data sets. In addition, the k-means algorithm relies on a distance metric, usually the Euclidean distance, which is not suitable for time series. It may not be the best choice for data, because the time series may have different scaling and offsets on the time axis; therefore, this embodiment first converts the original time series data into feature vectors that are easier to process and compare based on feature extraction, thereby reducing the computational burden of clustering directly on the original time series. In this way, the application more accurately captures the similarities between time series and improves the efficiency and effectiveness of clustering. Clustering based on these features can improve the accuracy and computational efficiency of clustering, making the clustering results more accurate and reliable, which is particularly suitable for large-scale time series data sets, and can significantly reduce computing time and resource consumption, while providing higher quality clustering results;

[0076] Based on the cluster identifier, users are divided into k groups, and the total load after summing up the power load of each group of users is modeled and predicted. The model framework can use ARIMA, LSTM, CNN-LSTM-Attention and CNN-GRU-Attention model frameworks. Based on the characteristics of each model framework, groups can be freely selected according to specific application scenarios. This application does not impose any restrictions on this;

[0077] ARIMA model: It is suitable for stationary time series and can better capture long-term and short-term changes, but the data needs to meet the stationarity assumption;

[0078] LSTM model: suitable for non-stationary time series and can handle long-term dependencies, but it consumes large computing resources and takes a long time to train;

[0079] CNN-LSTM-Attention and CNN-GRU-Attention: They combine the advantages of convolutional neural networks (CNN) and long short-term memory networks (LSTM or GRU), and use the attention mechanism to increase the focus on important features. They are suitable for multi-feature load forecasting and have good forecasting effects, but the model complexity is relatively high.

[0080] Embodiment 2

[0081] Figure 2is a system schematic diagram of a clustering-based industry segmentation electricity sales forecasting device according to another exemplary embodiment, the device comprising:

[0082] Historical data acquisition module 1: used to acquire historical power consumption data of each power user; based on the historical power consumption data, acquire the power consumption data of each power user in a historical preset time period;

[0083] The power consumption behavior characterization vector acquisition module 2 is used to acquire the power consumption behavior characterization vector of each power user in the historical preset period based on the power consumption data of each power user in the historical preset period;

[0084] Attribute vector acquisition module 3: used to acquire the attributes of each power user, vectorize the attributes of each power user, and obtain the attribute vector of each power user;

[0085] Characterization vector acquisition module 4: used to concatenate the power consumption behavior characterization vector of each power user and the corresponding self-attribute vector to obtain the characterization vector of each power user;

[0086] Clustering module 5: used to cluster each power user based on the characterization vector of each power user using a k-means clustering method to obtain a cluster identifier corresponding to each power user;

[0087] Group prediction module 6: It is used to group the power users according to the cluster identifier corresponding to each power user, sum the power load of the power users in the same group, iteratively train the pre-built model framework with the total load obtained by summing up each group of power users, and obtain the prediction model corresponding to each group of power users, and predict the power load of each power user in the corresponding group through the prediction model corresponding to each group of power users.

[0088] Embodiment three:

[0089] This embodiment provides a storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a main controller, each step in the above method is implemented;

[0090] It is understandable that the storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0091] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.

[0092] It should be noted that, in the description of the present invention, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise specified, the meaning of "plurality" refers to at least two.

[0093] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0094] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0095] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0096] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0097] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0098] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0099] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. The clustering-based industry segmentation electricity sales forecasting method is characterized by: The method comprises: Acquire historical electricity consumption data of each electricity user; acquire electricity consumption data of each electricity user in a historical preset time period based on the historical electricity consumption data; Based on the electricity consumption data of each electricity user in the historical preset time period, obtaining the electricity consumption behavior characterization vector of each electricity user in the historical preset time period; Acquire the attributes of each power user, and vectorize the attributes of each power user to obtain the attribute vector of each power user; The power consumption behavior characterization vector of each power user and the corresponding self-attribute vector are concatenated to obtain the characterization vector of each power user; Based on the characterization vectors of the power users, the power users are clustered using a k-means clustering method to obtain a cluster identifier corresponding to each power user; The power users are grouped according to their corresponding cluster identifiers, and the power loads of the power users in the same group are summed up. The pre-built model framework is iteratively trained with the total load obtained by summing up each group of power users to obtain the prediction model corresponding to each group of power users. The power loads of each power user in the corresponding group are predicted using the prediction model corresponding to each group of power users.

2. The method according to claim 1, characterized in that The step of obtaining the power consumption behavior characterization vector of each power user in the historical preset period based on the power consumption data of each power user in the historical preset period includes: Setting a window size of the window function, and rolling acquiring a window representation vector in the power consumption data of each power user in a preset historical period based on the preset window function; Obtaining a peak-valley value time period vector based on the power consumption data of each power user in a historical preset time period; The window characterization vectors and the peak-valley value time period vectors of each power user are spliced ​​to obtain the power consumption behavior characterization vectors of each power user in the historical preset time period.

3. The method according to claim 2, characterized in that The step of rolling and acquiring the window characterization vector in the power consumption data of each power user in a preset historical period based on a preset window function includes: Based on the preset window function, rolling obtain the average power consumption in each window in the historical preset time period, and use the average power consumption as the average value representation vector; Based on the preset window function, the maximum power consumption in each window in the historical preset time period is obtained in a rolling manner, and the maximum power consumption is used as a maximum value characterization vector; Based on the preset window function, the minimum power consumption in each window in the historical preset time period is obtained in a rolling manner, and the minimum power consumption is used as a minimum value characterization vector; The average value characterization vector, the maximum value characterization vector, and the minimum value characterization vector are concatenated to obtain a window characterization vector in the electricity consumption data of each power user in a historical preset time period.

4. The method according to claim 3, characterized in that The step of obtaining the peak-valley value time period vector based on the power consumption data of each power user in the historical preset time period includes: Divide a day into n time periods in units of x hours, obtain the peak power consumption and valley power consumption of each power user in the historical preset time period, and obtain the number of times the peak power consumption and valley power consumption of each power user occur in the corresponding time period, and obtain the peak time period vector and valley time period vector respectively; The peak time period vector and the valley time period vector of each power user are spliced ​​together to obtain the peak and valley time period vector of the power consumption data of each power user in the historical preset time period.

5. The method according to claim 4, characterized in that The attributes of each power user are the industry classification of each power user. The industry classification of each power user is vectorized to obtain the attribute vector of each power user.

6. The method according to claim 5, characterized in that The electricity consumption data of the historical preset time period is the electricity consumption data within Y days from the day to be predicted.

7. The clustering-based industry segmentation electricity sales prediction device is characterized by: The device comprises: Historical data acquisition module: used to acquire historical power consumption data of each power user; based on the historical power consumption data, acquire the power consumption data of each power user in a historical preset time period; An electricity consumption behavior characterization vector acquisition module: used to acquire the electricity consumption behavior characterization vector of each electricity user in a historical preset period based on the electricity consumption data of each electricity user in a historical preset period; Attribute vector acquisition module: used to acquire the attributes of each power user, vectorize the attributes of each power user, and obtain the attribute vector of each power user; A characterization vector acquisition module is used to concatenate the characterization vector of the power consumption behavior of each power user and the corresponding self-attribute vector to obtain the characterization vector of each power user; Clustering module: used to cluster each power user based on the characterization vector of each power user using the k-means clustering method to obtain a cluster identifier corresponding to each power user; Group prediction module: It is used to group the power users according to the cluster identifier corresponding to each power user, sum the power load of the power users in the same group, iteratively train the pre-built model framework with the total load obtained by summing up each group of power users, and obtain the prediction model corresponding to each group of power users, and predict the power load of each power user in the corresponding group through the prediction model corresponding to each group of power users.

8. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the main controller, each step of the clustering-based industry-segmented electricity sales forecasting method as described in any one of claims 1 to 6 is implemented.

Images