User electricity price portrait construction method and system based on feature clustering

Abstract

The invention provides a user electricity price portrait construction method and system based on feature clustering, and relates to the field of user feature analysis, and the method comprises the steps: obtaining electricity consumption data and electricity price data of each power user in a power user set, extracting the electricity price feature of each power user from the electricity consumption data and the electricity price data, and forming a feature data set; on the basis of the preprocessed feature data set, grouping a power consumer set by using a K-means clustering algorithm; and taking the clustering center of each group as a typical user of the group, carrying out portrait construction on each group according to feature data of the typical user, obtaining a multi-dimensional information label of each group, and taking the multi-dimensional information label as an electricity price portrait of the power user in the group. According to the method, the multi-dimensional feature information is combined, the user electricity price portrait is generated, and market-oriented user electricity consumption transaction behaviors can be accurately analyzed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of user feature analysis, in particular to a user electricity price portrait construction method and system based on feature clustering. BACKGROUND

[0002] With the deepening of global energy transformation and power market reform, the traditional power management mode is gradually changing to a diversified and competitive market mechanism. Electricity price, as the core signal of power market resource allocation, is the key to power system reform. Its formation mechanism has shifted to a dynamic electricity price adjustment mode through bilateral negotiation, centralized bidding, and listing, further enhancing the complexity and flexibility of electricity prices. This requires power companies to continuously improve the lean degree of electricity price management. By mining the electricity consumption patterns and transaction behavior of users, combined with user demand and electricity consumption characteristics, it can convert massive and chaotic user electricity consumption data into typical portraits with clear behavior characteristics in a complex and changing market environment, enabling power companies to systematically grasp the electricity consumption patterns, price sensitivity, and market interaction patterns of different user groups, and further enabling the accurate identification and positioning of abnormal electricity prices.

[0003] User electricity price is closely related to their electricity consumption behavior. Accurate analysis of electricity consumption behavior patterns can provide a basis for electricity price classification. Currently, scholars at home and abroad have conducted in-depth research on the construction of electricity consumption behavior portraits of power users, which can be divided into three steps: electricity load pattern extraction, customer electricity consumption feature selection, and behavior feature quantification. However, most of the research on market-oriented electricity user portraits does not consider the transaction behavior of users, and the research on electricity prices is mostly focused on market-oriented electricity prices. SUMMARY

[0004] To solve the above problems, the present application proposes a user electricity price portrait construction method and system based on feature clustering, which generates multiple typical user electricity price portraits by combining multi-dimensional feature information and analyzing market-oriented user electricity transaction behavior.

[0005] According to some embodiments, the present application adopts the following technical solutions: A user electricity price portrait construction method based on feature clustering, comprising: Obtaining electricity consumption data and electricity price data of each power user in a power user set, extracting electricity price features of each power user from the data, and forming a feature data set; Based on the pre-processed feature data set, using the K-means clustering algorithm to group the power user set; Taking the clustering center of each group as the typical user of the group, constructing a portrait for each group based on the feature data of the typical user, and obtaining the multi-dimensional information label of each group as the electricity price portrait of the power users in the group.

[0006] According to some embodiments, the present application adopts the technical solutions as follows: A user electricity price portrait construction system based on feature clustering comprises: A data acquisition module is configured to acquire electricity consumption data and electricity price data of each electricity user in a set of electricity users, extract electricity price features of each electricity user therefrom, and form a feature data set; A user grouping module is configured to group the set of electricity users based on the preprocessed feature data set by using a K-means clustering algorithm; A portrait construction module is configured to take a clustering center of each group as a typical user of the group, construct a portrait of each group according to feature data of the typical user, and obtain a multi-dimensional information label of each group as an electricity price portrait of an electricity user in the group.

[0007] According to some embodiments, the present application adopts the technical solutions as follows: A computer program product comprises a computer program, which, when executed by a processor, implements the user electricity price portrait construction method based on feature clustering.

[0008] According to some embodiments, the present application adopts the technical solutions as follows: A non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the user electricity price portrait construction method based on feature clustering.

[0009] According to some embodiments, the present application adopts the technical solutions as follows: An electronic device comprises a processor, a memory and a computer program; the processor is connected with the memory, and the computer program is stored in the memory; when the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device implements the user electricity price portrait construction method based on feature clustering.

[0010] Compared with the prior art, the present application has the following beneficial effects: The present application extracts electricity price features from electricity consumption behavior and transaction behavior, and uses a K-means clustering algorithm to construct a typical electricity price classification model, effectively capturing the random volatility, complex correlation and regional difference of electricity prices; on this basis, by deeply analyzing the electricity consumption behavior trajectory of a typical user in the whole cycle, a representative behavior label is generated, forming a typical user electricity price portrait with complete structure and distinctive features, thereby deeply revealing the characteristic mode, preference tendency and dynamic change law of user electricity consumption behavior. BRIEF DESCRIPTION OF DRAWINGS

[0011] The accompanying drawings, which form a part of this specification, are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification. The embodiments of the application, and their

[0012] Figure 1 is a feature clustering-based user electricity price portrait construction method flowchart of embodiment 1. Figure 2 is the principal component contribution rate and the cumulative contribution rate of principal components in embodiment 1. Figure 3 is the principal component score after dimension reduction in embodiment 1. Figure 4 is an industry 1 clustering evaluation index result graph in embodiment 1. Figure 5 is an industry 2 clustering evaluation index result graph in embodiment 1. Figure 6 is an industry 1 type A feature and electricity quantity curve graph in embodiment 1, wherein figure 6(a) is a feature curve and figure 6(b) is a daily electricity consumption curve. Figure 7 is an industry 1 type B feature and electricity quantity curve graph in embodiment 1, wherein figure 7(a) is a feature curve and figure 7(b) is a daily electricity consumption curve. Figure 8 is a feature portrait graph of a typical user in the chemical industry in embodiment 1. DETAILED DESCRIPTION

[0013] The application will be further described below with reference to the drawings and embodiments.

[0014] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the application. Unless otherwise indicated, all technical and scientific terms used herein have the same meaning as would be commonly understood by one of ordinary skill in the art to which the application pertains.

[0015] It should be noted that the terms used herein are only intended to describe specific embodiments and are not intended to limit exemplary embodiments according to the present application. As used herein, the singular form is intended to include the plural form unless the context clearly indicates otherwise, and it should be understood that when the terms "comprise" and / or "comprises" are used in this specification, they mean that the features, steps, operations, devices, components and / or combinations thereof are present.

[0016] Embodiment 1 In an embodiment of the present application, a feature clustering-based user electricity price portrait construction method is provided, which generates a user electricity price portrait by combining multi-dimensional feature information, can accurately analyze market-oriented user electricity transaction behavior, and specifically includes: Step S1: Obtain the power consumption data and electricity price data of each power user in the power user set, extract the electricity price characteristics of each power user, and form a feature data set.

[0017] At present, the electricity market rules stipulate that the electricity price of power users is composed of on-grid electricity price, on-grid link loss cost, transmission and distribution price (including cross-subsidy), system operation cost (including coal capacity electricity cost, auxiliary service cost, pumped storage capacity electricity cost, etc.), government fund and additional cost. Among them, the user side subject spot settlement price in the on-grid electricity price is calculated by the hourly node electricity price, which is determined by the weighted average of 15-minute clearing electricity price and clearing electricity quantity of each node on the electricity consumption side. Since the transmission and distribution price, system operation cost, government fund and additional cost are relatively stable, in general, the abnormal electricity price of power users is mainly caused by marketized transaction electricity price.

[0018] According to the disclosure data of a provincial power trading center in a certain month, the wholesale users only account for 4% of the user side users (power selling companies), and direct participation in wholesale trading has certain professional technical threshold requirements. Therefore, the wholesale users are not considered in this embodiment, and the feature data set is constructed based on the user power consumption data and electricity price data, the industry and user power consumption behavior and transaction behavior are analyzed, and the feature data set X is represented by the formula: (1) (2) (3) (4) (5) (6) (7) (8) In the formula, X includes 10 features, specifically: , , , are the mean, variance, maximum and minimum of the user power, n is the number of valid power consumption days in the current month, is the daily power consumption of the user; , , , are the average peak power consumption, average valley power consumption, peak power consumption ratio and valley power consumption ratio, , are the total peak power consumption and total valley power consumption of the user, , are the number of peak and valley hours in the current month. This represents the monthly arithmetic average of the day-ahead settlement prices on the user side of the spot market. The average monthly market-based electricity price for users, where s, , , These are the number of time periods, the current time period's electricity consumption, the total time period's electricity consumption, and the current time period's market-based electricity price.

[0019] Step S2: Based on the preprocessed feature dataset, the set of electricity users is grouped using the K-means clustering algorithm.

[0020] To comprehensively analyze user behavior patterns, the extracted electricity price features inevitably suffer from feature redundancy, greatly increasing the complexity of clustering solutions. Principal Component Analysis (PCA) is a linear, unsupervised, and global data dimensionality reduction algorithm. By using PCA to reduce the dimensionality of the input feature dataset, clustering accuracy can be effectively improved.

[0021] PCA uses the correlation between variables in the data to reduce dimensionality, and uses the principal components found in the data to represent the original data. The specific steps are as follows: (1) Data centering: The feature dataset X is centered, that is, the mean of each feature is calculated. This can be expressed as a formula: (9) Where m is the number of user samples in the feature dataset X. Let n be the number of features in the feature dataset X. In this embodiment, n=10.

[0022] (2) Calculate the covariance matrix: Calculate the covariance matrix C of the decentralized matrix E (dimension m×n). The calculation formula is as follows: (10) (11) in, These are the elements in the decentralized matrix E.

[0023] (3) Eigenvalue decomposition: The eigenvectors of the covariance matrix are obtained by solving the following equations. and the corresponding eigenvalues : (12) (4) Principal component selection: Arrange the eigenvalues ​​in descending column order from left to right, while maintaining the correspondence between the eigenvectors and the sorted eigenvalues. Select the eigenvectors corresponding to the k largest eigenvalues ​​to form matrix W.

[0024] (5) Data projection: Project the decentralized matrix E onto a new subspace, and calculate the output sample set Y after dimensionality reduction to k dimensions by Y=EW.

[0025] Principal component contribution rate and cumulative contribution rate are as follows Figure 2 As shown, the principal component scores after dimensionality reduction are as follows: Figure 3 As shown, the first four components are selected as the output feature dimensions. After PCA processing, the dimensionality of the dataset is reduced while containing most of the original information, and their cumulative contribution rate exceeds 98%.

[0026] The electricity spot market is affected by the coupled disturbances from multiple influencing factors. Relying solely on data-driven methods or human experience to carry out electricity price management and anomaly identification will face significant implementation bottlenecks. By deeply analyzing the user's full-cycle behavior trajectory and combining it with multi-dimensional feature clustering to construct a dynamically iterative user electricity price profile, we can overcome the inherent defects of data-driven approaches and achieve a key leap from "batch data screening" to "precise individual insights".

[0027] K-means is a widely used clustering algorithm that iteratively optimizes the location of cluster centers by minimizing the Within-Cluster Sum of Squares (WCSS), grouping data points that are as similar as possible into K clusters. The within-cluster sum of squares J is shown in the formula, where K is the number of clusters. Let y be the point set of the i-th cluster, and y be the set of points belonging to the i-th cluster. The data point, i.e., a user sample in the sample set Y, It is the centroid of the i-th cluster. Represents the relationship between data point y and cluster center The square of the Euclidean distance between them: (13) K-means converges gradually by iteratively adjusting the cluster assignments to minimize the distance between data points within a cluster and the centroid. The specific steps are as follows: (1) Initialization: K data points are randomly selected as the initial cluster centers, which are called cluster centers. , ,…, The first K cluster centers are represented by the superscript (0), which indicates the 0th iteration (initial state).

[0028] (2) Assignment: Using the Euclidean distance d, each data point in the dataset is assigned to the nearest cluster center. Where n is the dimension of the data, and the Euclidean distance d is expressed by the formula: (14) The cluster allocation rules are as follows: (15) (3) Update and iteration: Recalculate the center of each cluster, and repeat the steps of assignment and update iteration until the cluster center no longer changes or reaches the preset number of iterations. The center calculation formula is: (16) Step S3: Take the clustering center of each group as the typical user of this group. According to the characteristic data of the typical user, construct a portrait for each group to obtain multi-dimensional information labels for each group, and use them as the electricity price portraits of the electricity users within the group.

[0029] In the electricity spot market, the electricity consumption behaviors of industrial and commercial users are directly bound to production and operation activities. The electricity consumption behaviors of users are the key and fundamental links in the formation of electricity prices. From the core dimensions of industry anchoring, electricity consumption behaviors, and market interaction, deeply disassemble the full-cycle behavior trajectories of users, and design a multi-dimensional user electricity price portrait system, specifically as follows: (1) Dimension of electricity consumption behavior characteristics It includes three types of labels: "electricity consumption period", "electricity consumption scale", and "periodic fluctuation": For the electricity consumption period label, according to the peak electricity consumption ratio and the valley electricity consumption ratio, it is divided into peak-valley type (peak electricity consumption ratio > 60% and valley electricity consumption ratio < 30%), anti-peak-valley type (peak electricity consumption ratio < 30% and valley electricity consumption ratio > 60%), random type (including two cases: peak electricity consumption ratio > 60% and valley electricity consumption ratio > 30%, peak electricity consumption ratio < 60% and valley electricity consumption ratio < 30%), and balanced type (other users except the above users).

[0030] For the electricity consumption scale label, according to the electricity consumption, it is divided into micro users, small users, medium users, etc.; using the P95 and P75 percentile methods, calculate the monthly total electricity consumption data of all industry users. Users with electricity consumption ranking in the top 5% (monthly total electricity consumption ≥ P95) are super-large users, users with electricity consumption ranking in the top 5% - 25% (monthly total electricity consumption ≥ P75 and < P95) are large users, users with electricity consumption ranking in the middle 50% (monthly total electricity consumption ≥ P25 and < P75) are medium users, users with electricity consumption ranking in the bottom 5% - 25% (monthly total electricity consumption ≥ P5 and < P25) are small users, and users with electricity consumption ranking in the bottom 5% (monthly total electricity consumption < P5) are micro users.

[0031] For the periodic fluctuation label, according to the mutation fluctuation coefficient, it is divided into stable fluctuation type (mutation fluctuation coefficient ≤ 0.3) and mutation fluctuation type (mutation fluctuation coefficient ≥ 0.6), and moderate volatility (0.3 < mutation fluctuation coefficient < 0.6); among them, the formula for the mutation fluctuation coefficient MU is: (17) in, , These represent the average and variance of user battery usage, respectively.

[0032] (2) Industry and production characteristics dimension The system includes two tags: "Industry Classification" and "Production Cycle." Industry classifications are categorized into steel, coal, rubber, hospitals, oil extraction, etc. These tags are user-defined attribute data and do not require calculation. The production cycle, based on the fluctuation coefficient and electricity price sensitivity coefficient, is divided into fixed cycle (fluctuation coefficient ≤ 0.3), order-driven elastic cycle (fluctuation coefficient > 0.6 and absolute value of electricity price sensitivity coefficient < 0.6), and market response elastic cycle (fluctuation coefficient > 0.6 and absolute value of electricity price sensitivity coefficient ≥ 0.6). The electricity price sensitivity coefficient... The sensitivity between user electricity consumption behavior and electricity prices is defined by the following formula: (17) In the formula, It is the user's average load in month t; This represents the average load of a certain user in the industry in month t-1. This is the electricity price for a certain user in the industry in month t. This is the electricity price for a certain user in the industry in month t-1.

[0033] (3) Market interaction characteristics The "Market Response Behavior" label is based on the mutation volatility coefficient and the electricity price sensitivity coefficient, and is divided into high electricity price sensitivity type (absolute value of electricity price sensitivity coefficient > 0.7), medium electricity price sensitivity type (0.3 < absolute value of electricity price sensitivity coefficient ≤ 0.7), and low electricity price sensitivity type (absolute value of electricity price sensitivity coefficient ≤ 0.3).

[0034] This embodiment selects actual data from market-oriented retail users in a certain region as the research sample for in-depth analysis. To verify the effectiveness of the K-means clustering proposed in this embodiment, fuzzy C-means clustering is introduced as a supplementary control model M2 for comparative analysis.

[0035] To effectively determine the number of typical user electricity price types in different industries, a profile coefficient was selected. Silhouette Coefficient, Davies-Bouldin Index (DBI), Gap Statistic (Gap), Calinski-Harabaz Index (CHI) Four metrics are used to evaluate the classification effectiveness of the clustering model: (18) (19) (20) (twenty one) In the formula, , respectively, are the average distance between sample point i and all points in the nearest other clusters, and the average distance between i and all other points in the same cluster. , , , , , These represent the number of clusters, the centroid of cluster i, the centroid of cluster j, the average distance from the centroid to the midpoint of cluster i, the average distance from the centroid to the midpoint of cluster j, and the distance between the centroids of cluster i and cluster j, respectively.

[0036] To enhance the generalizability and validity of the research, cluster analysis was conducted using user groups from two different industries as sample data. The evaluation results of the number of clusters are shown in Table 1.

[0037] Table 1 Evaluation index values ​​of each model

[0038] Table 1 shows that the overall performance of the M2 model is lower than that of the M1 model under different numbers of clusters. Although the fuzzy C-means model has its inherent advantages in handling data fuzziness, allowing the same data point to belong to all clusters with different membership degrees, based on the comparison results of four evaluation indicators under different industries, M2 only outperforms the mean model M1 in terms of Gap coefficient. It is evident that the soft assignment strategy of fuzzy C-means clustering does not offer a significant advantage in this regard and may even introduce unnecessary computational complexity. Considering the comprehensive evaluation indicator calculation results, the method proposed in this paper can more efficiently and accurately characterize the true data distribution. The evaluation indicator results for the number of clusters are as follows: Figures 4-5 As shown, considering the results of the four evaluation indicators, the optimal number of clusters in industry 1 is 3; the optimal number of clusters in industry 2 is 5.

[0039] To further demonstrate the classification effect of the clustering model, a random user under two different feature types in Industry 1 was selected for analysis. The feature curves and the daily electricity consumption curves for a certain month are shown in Figures 6 and 7.

[0040] In the field of electricity market user characteristics research, users from different industries, with varying electricity consumption behaviors, and under different market interactions exhibit differentiated characteristics. To more clearly demonstrate the comprehensive characteristics of a typical user group, this paper constructs a user characteristic system combining three dimensions: industry and production, electricity consumption behavior, and market interaction. Figure 8The image presents a characteristic profile of a typical user in the chemical industry. This user belongs to the chemical industry, characterized by a fixed production cycle, random electricity consumption periods, a small user scale, and stable electricity consumption cycle fluctuations. In terms of market interaction, this user's electricity load lacks price elasticity, and its electricity consumption behavior is not sensitive to changes in electricity prices.

[0041] This embodiment proposes a user electricity price profiling method based on feature clustering. It extracts electricity price features from both electricity consumption behavior and transaction behavior, analyzes user electricity price features, constructs an electricity price profile, and analyzes the random fluctuations, complex correlations, and regional differences in electricity prices to gain a deeper understanding of the characteristics, preferences, and changing patterns of user electricity consumption behavior.

[0042] Example 2 One embodiment of the present invention provides a user electricity price profile construction system based on feature clustering, comprising: The data acquisition module is configured to: acquire the electricity consumption data and electricity price data of each electricity user in the electricity user set, extract the electricity price characteristics of each electricity user, and form a feature dataset; The user grouping module is configured to group the set of electricity users based on the preprocessed feature dataset using the K-means clustering algorithm; The profile building module is configured to: use the cluster center of each group as the typical user of the group, build a profile for each group based on the feature data of the typical user, obtain multi-dimensional information tags for each group, and use them as the electricity price profile of the electricity users in the group.

[0043] Example 3 One embodiment of the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned method for constructing a user electricity price profile based on feature clustering.

[0044] Example 4 In one embodiment of the present invention, a non-transitory computer-readable storage medium is provided for storing computer instructions. When the computer instructions are executed by a processor, they implement the aforementioned method for constructing a user electricity price profile based on feature clustering.

[0045] Example 5 One embodiment of the present invention provides an electronic device, including: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes the aforementioned method for constructing a user electricity price profile based on feature clustering.

[0046] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0047] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0048] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A user electricity price portrait construction method based on feature clustering, characterized in that, The method comprises the following steps: obtaining power consumption data and electricity price data of each power user in a power user set, and extracting electricity price features of each power user from the data, forming a feature data set, grouping the power user set based on the preprocessed feature data set by using a K-means clustering algorithm, taking the clustering center of each group as a typical user of the group, and constructing a portrait of each group according to the feature data of the typical user to obtain a multi-dimensional information label of each group, which is used as the electricity price portrait of the power users in the group. 2.The method of claim 1, wherein, The electricity price data includes online electricity price, online link line loss cost, power transmission and distribution price, system operation cost, and government fund. 3.The method of claim 1, wherein, The electricity price features include user power consumption mean, variance, maximum value, minimum value, peak average power consumption, valley average power consumption, peak power consumption ratio, valley power consumption ratio, monthly arithmetic mean of spot market user-side day-ahead settlement price, and monthly average market transaction electricity price of the user. 4.The method of claim 1, wherein, The preprocessing includes dimension reduction of the correlation of the features in the feature data set by using principal component analysis.

5. The user tariff profiling method based on feature clustering as claimed in claim 1, wherein, The grouping of the power user set by using the K-means clustering algorithm is to take the electricity price features of the power users as data points, to optimize the position of the cluster center through iteration, to minimize the total sum of intra-cluster squared error, and to group the data points into K clusters.

6. The user tariff profiling method based on feature clustering as claimed in claim 1, wherein, The multi-dimensional information label is to construct a user electricity price portrait system by combining three dimensions of power consumption behavior features, industry and production features, and market interaction features, and to perform multi-dimensional classification of the user groups by using the user electricity price portrait system. 7.A system for constructing a user tariff profile based on feature clustering, the system comprising: The method comprises the following steps: a data acquisition module configured to obtain power consumption data and electricity price data of each power user in a power user set, and to extract electricity price features of each power user from the data, forming a feature data set, a user grouping module configured to group the power user set based on the preprocessed feature data set by using a K-means clustering algorithm, a portrait construction module configured to take the clustering center of each group as a typical user of the group, and to construct a portrait of each group according to the feature data of the typical user to obtain a multi-dimensional information label of each group, which is used as the electricity price portrait of the power users in the group.

8. A computer program product comprising a computer program, characterized in that, The computer program is executed by a processor to implement the feature clustering-based user electricity price portrait construction method of any one of claims 1-6.

9. A non-transitory computer-readable storage medium, comprising: The non-transitory computer readable storage medium is used to store computer instructions, which are executed by a processor to implement the feature clustering-based user electricity price portrait construction method of any one of claims 1-6.

10. An electronic device, comprising: The electronic device comprises a processor, a memory, and a computer program; the processor is connected with the memory, and the computer program is stored in the memory; when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the feature clustering-based user electricity price portrait construction method of any one of claims 1-6. ​

Images