Electric charge recovery risk prediction method and device, electronic equipment and storage medium

Abstract

The invention provides an electricity charge recovery risk prediction method and device, electronic equipment and a storage medium, and the method comprises the steps: firstly extracting an electricity utilization feature corresponding to a first electricity utilization curve and an electricity charge payment feature corresponding to first electricity charge payment data; then, according to the electricity utilization characteristic and the electricity charge payment characteristic, determining an electricity charge behavior characteristic of the target user, then obtaining a first basic information characteristic of the target user, and inputting the electricity charge behavior characteristic of the target user and the first basic information characteristic into a risk prediction model obtained by training; according to the method and the device, the electric charge recovery risk corresponding to the target user is obtained, the electric charge recovery risk prediction is performed through the risk prediction model obtained through training, and the accuracy and the efficiency of electric charge recovery risk prediction are further improved.

Description

Technical Field

[0001] The present application relates to the technical field of prediction of electricity fee recovery risk, and in particular to a method, device, electronic device and storage medium for predicting electricity fee recovery risk. Background Art

[0002] Electricity fee recovery has always been a key focus of power supply companies. In order to prevent the risk of electricity fee recovery, power supply companies need to predict the risk of electricity fee recovery for electricity users in advance. With the development of smart grids, power companies are faced with the problem of collecting and analyzing a large amount of user electricity consumption data. In current related technologies, the risk estimation and judgment of electricity fee recovery mainly rely on manual experience, which is highly subjective and inefficient. Summary of the invention

[0003] In view of this, the purpose of this application is to propose a method, device, electronic device and storage medium for predicting electricity fee recovery risk to solve or partially solve the above technical problems.

[0004] Based on the above purpose, this application provides a method for predicting electricity fee recovery risk, including:

[0005] Obtaining a first power consumption curve of a target user to be predicted within a preset time range, and extracting power consumption characteristics corresponding to the first power consumption curve;

[0006] Acquire first electricity fee payment data of the target user within the preset time range, and extract electricity fee payment features corresponding to the first electricity fee payment data;

[0007] Determining the electricity fee behavior characteristics of the target user based on the electricity usage characteristics and the electricity fee payment characteristics;

[0008] Acquire a first basic information feature of the target user;

[0009] The electricity charge behavior characteristics of the target user and the first basic information characteristics are input into the risk prediction model obtained through training to obtain the electricity charge recovery risk corresponding to the target user.

[0010] Based on the same concept, the present application also proposes a prediction device for electricity fee recovery risk, including:

[0011] A first acquisition module, acquiring a first power consumption curve of a target user to be predicted within a preset time range, and extracting power consumption features corresponding to the first power consumption curve;

[0012] A second acquisition module, acquiring first electricity fee payment data of the target user within the preset time range, and extracting electricity fee payment features corresponding to the first electricity fee payment data;

[0013] A determination module, which determines the electricity fee behavior characteristics of the target user based on the electricity usage characteristics and the electricity fee payment characteristics;

[0014] A third acquisition module is used to acquire the first basic information feature of the target user;

[0015] The prediction module inputs the electricity charge behavior characteristics of the target user and the first basic information characteristics into the risk prediction model obtained through training to obtain the electricity charge recovery risk corresponding to the target user.

[0016] Based on the same concept, the present application also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for predicting the risk of electricity fee recovery as described above is implemented.

[0017] Based on the same concept, the present application also proposes a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to execute the above-mentioned method for predicting electricity fee recovery risks.

[0018] Based on the same concept, the present application also proposes a computer program product, which includes a computer program, and the computer program is executed by one or more processors so that the processors execute the method for predicting electricity fee recovery risks as described above.

[0019] As can be seen from the above, the prediction method, device, electronic device and storage medium for electricity bill recovery risk provided by the present application first obtain the first electricity consumption curve of the target user to be predicted within a preset time range, and extract the electricity consumption characteristics corresponding to the first electricity consumption curve, then obtain the first electricity bill payment data of the target user within the preset time range, and extract the electricity bill payment characteristics corresponding to the first electricity bill payment data; then determine the electricity bill behavior characteristics of the target user according to the electricity consumption characteristics and the electricity bill payment characteristics, then obtain the first basic information characteristics of the target user, and input the electricity bill behavior characteristics of the target user and the first basic information characteristics into the risk prediction model obtained by training to obtain the electricity bill recovery risk corresponding to the target user, and use the risk prediction model obtained by training to predict the electricity bill recovery risk, thereby further improving the accuracy and efficiency of the electricity bill recovery risk prediction. At the same time, by further processing the target user's electricity consumption curve and the first electricity bill payment data, the target user's electricity bill behavior characteristics are obtained. Compared with directly inputting the electricity consumption curve and the first electricity bill payment data into the risk prediction model, the electricity bill behavior characteristics are obtained, which reflects the user's long-term electricity bill payment behavior, and further improves the accuracy of the output result of the risk prediction model. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the present application or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 A schematic diagram of an application scenario of an embodiment of the present application;

[0022] Figure 2 A schematic diagram of a flow chart of a method for predicting electricity fee recovery risk according to an embodiment of the present application;

[0023] Figure 3 A schematic diagram of the structure of a device for predicting electricity fee recovery risk according to an embodiment of the present application;

[0024] Figure 4 A schematic diagram of a specific electronic device hardware structure of an embodiment of the present application. DETAILED DESCRIPTION

[0025] It is understandable that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and relevant provisions.

[0026] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should be the usual meanings understood by people with general skills in the field to which the present application belongs. The "first", "second" and similar words used in the present application do not represent any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements or objects appearing in front of the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connecting" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, which can be direct connections, indirect connections, wired connections, and wireless connections. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly. In addition, in the description of the present application, unless otherwise specified, the term "multiple" refers to two or more. The term "and / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects are in an "or" relationship.

[0027] The principles and spirit of the present application will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present application, and are not intended to limit the scope of the present application in any way. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0028] It is understandable that before using the technical solutions of each embodiment of the present application, the type, scope of use, usage scenarios, etc. of the personal information involved will be informed to the user in an appropriate manner, and the user's authorization will be obtained.

[0029] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly remind the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can independently choose whether to provide personal information to the electronic device, application, server, storage medium or other software or hardware that performs the operation of the technical solution of the present application according to the prompt message.

[0030] As an optional but non-limiting implementation, in response to receiving the user's active request, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0031] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation method of this application. Other methods that meet relevant laws and regulations may also be applied to the implementation method of this application.

[0032] It should be understood herein that any number of elements in the drawings is for illustration rather than limitation, and any naming is only for distinction rather than having any limiting meaning.

[0033] As mentioned in the background technology section, the current related technologies mainly rely on manual experience to estimate and judge the risk of electricity fee recovery. This method is highly subjective and inefficient. For example, in some related technologies, various basic information of users is collected, such as user type, region, industry, electricity consumption level, account balance, account age, etc. Based on these characteristics, experts formulate scoring rules, that is, the risk score corresponding to each feature is summarized to obtain the user's total score. According to the set scoring threshold, users are divided into different risk levels, and high-risk users are tracked and recovered. This method relies heavily on manual subjective judgment, and is inefficient,

[0034] In summary, in order to solve the above problems, the present application embodiment proposes a method for predicting electricity fee recovery risk, including:

[0035] First, a first electricity consumption curve of the target user to be predicted within a preset time range is obtained, and the electricity consumption characteristics corresponding to the first electricity consumption curve are extracted. Then, the first electricity bill payment data of the target user within the preset time range is obtained, and the electricity bill payment characteristics corresponding to the first electricity bill payment data are extracted. Then, the electricity bill behavior characteristics of the target user are determined according to the electricity consumption characteristics and the electricity bill payment characteristics. Then, the first basic information characteristics of the target user are obtained, and the electricity bill behavior characteristics of the target user and the first basic information characteristics are input into the risk prediction model obtained through training to obtain the electricity bill recovery risk corresponding to the target user. The electricity bill recovery risk prediction is performed through the risk prediction model obtained through training, which further improves the accuracy and efficiency of the electricity bill recovery risk prediction. At the same time, the electricity bill behavior characteristics of the target user are obtained by further processing the electricity consumption curve and the first electricity bill payment data of the target user. Compared with directly inputting the electricity consumption curve and the electricity bill payment data into the risk prediction model, it can help the risk prediction model to more accurately confirm the target user's electricity bill payment mode, and further improve the accuracy of the output result of the risk prediction model.

[0036] In some specific application scenarios, the prediction method of the electricity fee recovery risk of the present application can be applied to various platforms or systems involving the prediction of electricity fee recovery risk. As an example, the application scenario includes at least one server and at least one terminal. Figure 1 , which is a schematic diagram of an application scenario of a method for predicting the risk of electricity fee recovery provided by an exemplary embodiment of the present disclosure. The application scenario includes a terminal device 101, a server 102, and a data storage system 103. Among them, the terminal device 101, the server 102 and the data storage system 103 can be connected through a wired or wireless communication network. The terminal device includes but is not limited to a desktop computer, a mobile phone, a mobile computer, a tablet computer, a media player, a smart wearable device, a personal digital assistant (PDA) or other electronic devices capable of realizing the above functions. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server and the terminal can communicate through the network to realize data transmission. Among them, the network can be a wired network or a wireless network, which is not specifically limited in this application.

[0037] The server can be a server that provides various services. Specifically, the server can be used to provide background services for applications running on the terminal. Optionally, in some implementations, the method for predicting the risk of electricity bill recovery provided in the embodiment of the present application can be executed by the server. Optionally, in some implementations, the method for predicting the risk of electricity bill recovery provided in the embodiment of the present application can be directly executed by the terminal device. The server can be hardware or software. When the server is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or it can be implemented as a single server. When the server is software, it can be implemented as multiple software or software modules (for example, software or software modules used to provide distributed services), or it can be implemented as a single software or software module. The embodiment of the present application does not specifically limit this.

[0038] Optionally, the wireless network or wired network uses standard communication technology and / or protocol. The network is usually the Internet, but it can also be any network, including but not limited to a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a dedicated network or any combination of a virtual private network). In some embodiments, the data exchanged through the network is represented by technologies and / or formats including hypertext mark-up language (HTML), extensible markup language (XML), etc. In addition, conventional encryption technologies such as secure socket layer (SSL), transport layer security (TLS), virtual private network (VPN), and Internet protocol security (IPsec) can be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can also be used to replace or supplement the above data communication technologies.

[0039] The data storage system stores a large amount of data. Optionally, the data storage system can partition the data according to different uses and data types. The source of the data includes but is not limited to an existing database, data crawled from the Internet, or data uploaded by the user when using the client.

[0040] The following describes the prediction method of the risk of electricity fee recovery according to the exemplary implementation of the present application in combination with a specific application scenario. It should be noted that the above application scenario is only shown to facilitate understanding of the spirit and principle of the present application, and the implementation of the present application is not limited in this regard. On the contrary, the implementation of the present application can be applied to any applicable scenario.

[0041] refer to Figure 2 , is a flow chart of a method for predicting electricity fee recovery risk according to an embodiment of the present application, the method comprising the following steps:

[0042] S101, obtaining a first power consumption curve of a target user to be predicted within a preset time range, and extracting power consumption features corresponding to the first power consumption curve.

[0043] In specific implementation, the target user to be predicted can be any electricity user, without limitation. The preset time range can be set as needed, for example, it can be set to a time range of nearly one month, nearly one year, etc. It should be noted that, in general, the time corresponding to the end point of the preset time range can be set to the day before the prediction. For example, the time for predicting the risk of electricity bill recovery is January 6, 2024, then the preset time range can be from January 5, 2023 to January 5, 2024. The first electricity consumption curve is the electricity consumption curve of the user within the preset time range. Optionally, the horizontal axis of the curve can be time, and the vertical axis can be electricity consumption. The extracted user features can be the number of users at each time point, or the historical average number of users, monthly electricity consumption, etc.

[0044] In some embodiments, extracting the power usage feature corresponding to the first power usage curve specifically includes:

[0045] Extracting key indicator values ​​of preset indicators corresponding to the first power consumption curve;

[0046] In response to determining that the key indicator value belongs to a preset value range, a key nonlinear feature corresponding to the preset value range is added to the first power consumption curve.

[0047] In the specific implementation, it is considered that if the basic information in the user curve of the user is directly relied on as the power consumption characteristics, it will not only fail to effectively reflect the user's power consumption behavior characteristics, but also fail to fully consider the key characteristics such as nonlinear patterns, steps, and periodicity in the power load curve. Therefore, in the embodiment of the present application, when extracting the power consumption characteristics corresponding to the first power consumption curve, not only some linear characteristics such as the number of users at different times and the average power consumption are extracted, but also the key nonlinear characteristics corresponding to the power consumption curve are further deeply excavated through the pre-set preset indicators and preset numerical ranges. Among them, the preset indicators and the corresponding preset numerical ranges can be set as needed. Optionally, in some embodiments, the preset indicators and the corresponding preset numerical ranges can be set according to the following Table 1.

[0048] Table 1

[0049]

[0050] For example, by calculating whether the electricity consumption in a certain month corresponding to the first electricity consumption curve has increased by more than 50% compared with the same historical period (the corresponding preset indicator at this time is the growth percentage of the current month's electricity consumption relative to the same historical period, and the corresponding preset value range is greater than 50%), the electricity consumption sudden increase feature is set to different values ​​to serve as the key nonlinear feature of the "electricity consumption sudden increase feature" in the first electricity consumption curve.

[0051] S102, obtaining first electricity fee payment data of the target user within the preset time range, and extracting electricity fee payment features corresponding to the first electricity fee payment data.

[0052] In specific implementation, the electricity bill payment characteristics include: average payment time interval, payment amount standard deviation, average electricity bill payment amount in the payment cycle, average payment interval duration, payment method of electricity bill (cash, bank transfer, electronic channel, etc.), number of payment channel changes, continuous payment duration, monthly electricity bill payment frequency factor, etc. Optionally, the user can select one or more characteristics from the above multiple electricity bill payment characteristics as the characteristics extracted from the first electricity bill payment data in the embodiment of the present application as needed.

[0053] S103: Determine the electricity fee behavior characteristics of the target user based on the electricity usage characteristics and the electricity fee payment characteristics.

[0054] In a specific implementation, after extracting the electricity consumption characteristics corresponding to the first electricity consumption curve and the electricity payment characteristics corresponding to the first electricity bill payment data, the electricity bill behavior characteristics of the target user are determined based on the electricity consumption characteristics and the electricity bill payment characteristics. Optionally, in some embodiments, the electricity consumption characteristics and the electricity bill payment characteristics can be input into a trained encoder, and then the encoder outputs the corresponding electricity bill behavior characteristics. Optionally, in some embodiments, multiple preset electricity bill behavior characteristics can be set first, and then based on the electricity consumption characteristics and the electricity bill payment characteristics, a target preset electricity bill behavior characteristic with the closest characteristic distance is determined from the multiple preset electricity bill behavior characteristics, and the target preset electricity bill behavior characteristic is determined as the electricity bill behavior characteristic of the target user.

[0055] In some embodiments, when determining the electricity bill behavior characteristics of the target user, the various electricity bill payment modes that occur in the user within a preset time range can be determined based on the electricity usage characteristics and the electricity bill payment characteristics, and then the frequency of each electricity bill payment mode occurring within a period of time (which can be set to 1 year) is counted, and the frequency of occurrence is quantified to construct the electricity bill behavior characteristics. Optionally, the electricity usage characteristics and electricity bill payment characteristics are generally extracted in multiples. The specific various electricity bill payment modes can be set as needed. In some embodiments, the various electricity bill payment modes include: one or more of an advance payment mode, an on-time payment mode, an intermittent payment mode, a delayed payment mode, an overdue payment mode, and an overdue power outage mode. It should be noted that after determining the electricity usage characteristics and the electricity bill payment characteristics, the electricity bill payment mode of the target user can be determined by the electricity usage characteristics and the electricity bill payment characteristics, and then the payment mode frequency within a certain period of time is quantified to construct the electricity bill behavior characteristics.

[0056] In some embodiments, determining the target user's electricity fee behavior characteristics based on the electricity usage characteristics and the electricity fee payment characteristics specifically includes:

[0057] Dividing the preset time range into a plurality of preset time periods;

[0058] Determine the principal component characteristics corresponding to the electricity consumption characteristics and the electricity fee payment characteristics in each preset time period by using the kernel principal component analysis technology;

[0059] Determine the electricity fee payment mode corresponding to each preset time period of the target user based on the principal component characteristics of each preset time period;

[0060] The electricity bill behavior characteristics of the target user are determined based on the electricity bill payment mode corresponding to each preset time period of the target user. In specific implementation, the preset time range can be divided into multiple preset time periods first, and the preset time periods can be set as needed when dividing, and there is no limitation on this. For example, if a preset time range is 1 year, the preset time range can be divided into 12 months, each month corresponds to a preset time period; if a preset time range is 1 month, the preset time range can be divided into 4 weeks, each week corresponds to a preset time period. After the preset time range is divided into multiple preset time periods, the principal component characteristics corresponding to the electricity consumption characteristics and the electricity bill payment characteristics in each preset time period are determined by the kernel principal component analysis technology (KPCA), and then the electricity bill payment mode corresponding to the target user in each preset time period is determined based on the principal component characteristics of each preset time period, and the electricity bill behavior characteristics of the target user are determined based on the electricity bill payment mode corresponding to the target user in each preset time period. Optionally, after determining the electricity bill payment mode corresponding to the target user in each preset time period, the electricity bill behavior characteristics of the target user can be determined by weighted summation. For example, a preset time range is 1 month, which is divided into 4 preset time periods per week. Among them, the electricity bill payment mode corresponding to 3 preset time periods is the delayed payment mode, and the electricity bill payment mode corresponding to one preset time period is the on-time payment mode. Then the final electricity bill behavior characteristics are three-quarters of the delayed payment mode and one-quarter of the on-time payment mode.

[0061] It should be noted that kernel principal component analysis is a nonlinear dimensionality reduction technique based on kernel techniques, which nonlinearly maps samples in the original space to the feature space F, and then performs principal component analysis in the feature space. In some embodiments, the kernel principal component analysis process includes: if the original payment behavior sequence data {x1, x2, ...x n}, select a kernel function k(x, y), such as a Gaussian RBF kernel, map the two samples to the inner product of the high-dimensional feature space, and calculate the kernel function value K between the samples ij = k(x i , x j ), construct the kernel matrix K, perform hierarchical kernel PCA, perform low-rank decomposition on K, extract the main low-frequency components, retain the most important p low-frequency principal components, and reduce the dimension. Use the low-frequency principal components as new features, calculate the kernel matrix again, implement PCA again, continue to extract fewer low-frequency components, repeat the iterative kernel PCA multiple times, retain fewer low-frequency components each time, and hierarchical iteration can extract the low-frequency and global information of the features, filter out high-frequency noise, and finally retain low-dimensional principal components that are sufficient to express nonlinear features. Optionally, when selecting a kernel function, the effects of different kernel functions and the number of principal components can be evaluated, and then the optimal mapping can be selected.

[0062] In some embodiments, determining the electricity fee payment mode corresponding to each preset time period of the target user based on the principal component characteristics of each preset time period specifically includes:

[0063] Obtain target features corresponding to each preset payment mode;

[0064] For each preset time period, the distance between the principal component feature and the target features corresponding to each preset payment mode is determined, and the preset payment mode corresponding to the target feature with the closest distance is determined as the electricity fee payment mode of the target user in the preset time period.

[0065] In specific implementation, when determining the electricity payment mode corresponding to each preset time period for the target user, the target features corresponding to each preset payment mode can be determined. For example, when determining the target features corresponding to the delayed payment mode, a large number of historical users with delayed payment can be collected first, and then the electricity payment features and electricity consumption features of these historical users can be extracted, and the principal component features can be generated as the target features based on the electricity payment features and electricity consumption features of the historical users. After determining the target features corresponding to each preset payment mode, matching can be performed based on the feature distance to determine the electricity payment mode for each preset time period.

[0066] In some embodiments, the various preset electricity fee payment modes include: advance payment mode, on-time payment mode, intermittent payment mode, delayed payment mode, overdue payment mode, overdue power outage mode, etc.

[0067] It should be noted that in addition to determining the electricity bill payment pattern of the target user in each preset time period through the above-mentioned embodiment, in some embodiments, the target probability can also be determined by other means. For example, the electricity bill payment pattern can be determined by a pre-trained neural network model, that is, the principal component features are input into the trained neural network model, and then the neural network model outputs the electricity bill payment pattern.

[0068] In some embodiments, in the process of training the risk prediction model, in order to accurately determine the target features corresponding to each preset payment mode. The electricity consumption sample features and electricity bill payment sample features of multiple sample users can be obtained first, and then the principal component features corresponding to the electricity consumption sample features and electricity bill payment sample features of each sample user can be determined by the kernel principal component analysis technology, and then the principal component features of each sample user are determined as a category to obtain a target category set; multiple category merging processes are performed so that the number of remaining categories after the category merging process is equal to the total number of preset electricity payment modes; wherein each category merging process includes: determining the distance between the principal component features corresponding to any two categories in the target category set, merging the two principal component features with the closest distance into a new category, and updating the target category set based on the new category.

[0069] In specific implementation, when clustering the principal component features of the sample users, the principal component features of each sample user are first regarded as a single cluster with a single element, and then the distance (dissimilarity) between the classes is calculated, the class with the shortest distance (i.e., the largest similarity) is merged, and the entire process is traversed to gradually merge the small classes until the end condition is reached (the number of remaining classes after the class merging process is equal to the total number of the electricity payment mode). Optionally, after the total number of classes equal to the preset electricity payment mode is obtained by merging the clusters, the corresponding preset electricity payment mode can be matched for each class according to the merged principal component features (target features) corresponding to each class. Optionally, when matching the corresponding preset electricity payment mode for each class, it can be determined which class of electricity payment mode the sample user corresponding to the merged principal component features belongs to. If there are multiple sample users, the number of sample users corresponding to each electricity payment mode can be counted, and then the electricity payment mode with the largest number can be determined as the preset electricity payment mode corresponding to the class. It should be noted that in some embodiments, the Euclidean distance can be used as the distance between the principal component features corresponding to two classes. Considering that the agglomerative hierarchical clustering algorithm based on Euclidean distance is greatly affected by noise points. When the distance between two classes is close, a small number of points with a close distance will be preferentially synthesized into a cluster, while most of the samples of the two classes are not actually close, resulting in clustering errors. The agglomerative hierarchical clustering algorithm based on Euclidean distance can be regarded as using a linear model to learn the decision boundary. Since it can only learn very simple linear decision boundaries, the algorithm is very sensitive to noise points and cannot effectively separate the categories. For situations that cannot be separated in linear space, the data can be increased in dimension, and the classification boundary can be found in the high-dimensional space, thereby avoiding the influence of noise points in the original space. Therefore, in some embodiments, the Gaussian kernel measurement method can be used to achieve the increase in dimensionality. Optionally, the distance between the principal component features corresponding to the two categories can be calculated by the following formula:

[0070]

[0071] Among them, x represents the principal component feature corresponding to one category, y represents the principal component feature corresponding to another category, and σ is the width parameter of the function, which controls the radial range of the function.

[0072] S104: Acquire the first basic information feature of the target user.

[0073] In specific implementation, the first basic information feature can be set as needed. Generally, some information related to the prediction of electricity fee recovery risk can be used as the first basic information feature as needed. For example, the first basic information feature may include: monthly user volume, average daily electricity consumption and other data information related to electricity consumption of the target user, as well as age of the target user, city level and other data information related to the user portrait.

[0074] In some embodiments, the first basic information feature of the target user may be determined by using Table 2.

[0075] Table 2

[0076]

[0077]

[0078] S105, inputting the electricity charge behavior characteristics of the target user and the first basic information characteristics into the risk prediction model obtained through training to obtain the electricity charge recovery risk corresponding to the target user.

[0079] In specific implementation, after determining the target user's electricity charge behavior characteristics and the first basic information characteristics of the target user, the two are input into the risk prediction model obtained by training to obtain the electricity charge recovery risk corresponding to the target user. It should be noted that the specific process of training the risk prediction model is not limited, and reference can be made to the training process of the neural network model in the relevant technology. Optionally, the electricity charge recovery risk can be set in advance according to needs. For example, the electricity charge recovery risk may include: high recovery risk, medium recovery risk, low recovery risk, etc., or the electricity charge recovery risk may include recovery risk and no recovery risk. Optionally, in some embodiments, in order to better input the target user's electricity charge behavior characteristics and the first basic information characteristics into the risk prediction model, the non-numerical information in the electricity charge behavior characteristics and the first basic information characteristics can be quantified first, so that the risk prediction model can better identify and process these characteristics. Optionally, in some embodiments, different electricity charge behavior characteristics can be numerically coded and quantified. For example, the numbers 1 and 3 can be used to represent two different electricity charge behavior characteristics respectively.

[0080] In some embodiments, the risk prediction model can be trained using a large amount of historical sample data. For example, the first basic information feature samples and electricity bill behavior feature samples corresponding to high recovery risk users can be input into a benchmark model corresponding to the risk prediction model, and then the parameters of the benchmark model can be adjusted to make its output consistent with the high recovery risk.

[0081] In some embodiments, the process of training the risk prediction model includes:

[0082] The parameters of the risk prediction model are adjusted by Monte Carlo tree search technology.

[0083] In specific implementation, Monte Carlo Tree Search (MCTS) is a tree data structure-based method that uses the MCTS algorithm to globally search for the optimal parameters of the risk prediction model, which is beneficial to improving the prediction performance.

[0084] In some embodiments, the process of adjusting the parameters of the risk prediction model by Monte Carlo tree search technology includes: in the selection phase of the Monte Carlo tree search algorithm, starting from the root node, making a decision by comparing the value of the candidate nodes, and selecting the child node with the greatest value. There are three types of candidate nodes in this phase, namely, fully expanded nodes, uncompleted expanded nodes and leaf nodes. In the expansion phase of the Monte Carlo tree search algorithm, according to the UCT (Upper Confidence Bound Apply to Tree, game tree search algorithm) value of its child nodes, a new node will be created in the decision tree as the new child node of the node after the current node is expanded. In the simulation phase of the Monte Carlo tree search algorithm, the basis for the current node movement selection is a random strategy: that is, a child node is randomly selected from the candidate child nodes and expanded until the last layer of nodes. In the back propagation phase of the Monte Carlo tree search algorithm, Q (v) propagates in the reverse direction along the decision tree to the root node, and the node state of the Monte Carlo simulation is changed to visited.

[0085] In some embodiments, before inputting the target user's electricity charge behavior characteristics and the first basic information characteristics into the trained risk prediction model, the method further includes:

[0086] In response to determining that the interval between a first time point corresponding to the end point of the preset time range and a second time point at which the risk prediction model completes training is greater than a preset time period, obtaining a second electricity usage curve, second electricity fee payment data, and second basic information characteristics of the target user from the second time point to the first time point;

[0087] The risk prediction model trained at the second time point is incrementally trained based on the electricity charge behavior characteristics obtained from the second electricity consumption curve and the second electricity charge payment data and the second basic information characteristics.

[0088] During specific implementation, in order to achieve periodic incremental optimization of the risk prediction model, the embodiment of the present application first determines whether the currently used risk prediction model is the most recently trained risk prediction model before inputting the target user's electricity bill behavior characteristics and the first basic information characteristics into the risk prediction model obtained through training, that is, the interval between the first time point corresponding to the end point of the preset time range and the second time point when the risk prediction model completes training is less than or equal to the preset time period. The preset time period can be set as needed and is not limited to this. For example, the preset time period can be set to one week. If the currently used risk prediction model needs to be optimized based on the newly generated data, the risk prediction model that has completed training at the second time point is incrementally trained based on the electricity bill behavior characteristics and the second basic information characteristics obtained from the second electricity consumption curve and the second electricity bill payment data.

[0089] In some embodiments, the process of incremental training of the risk prediction model includes: constructing an initial model, using historical electricity bill user data to train an initial risk prediction model as a baseline model for incremental learning, defining an incremental training cycle, for example, every Sunday can be used as an incremental training time point, in each cycle, new user data and feedback are collected as incremental training data, based on the initial model, incremental data are used for training to generate a new version of the risk prediction model. The training method is gradient based, the model is updated with the initial parameters as the starting point, and the effects of the new and old models are evaluated on the validation set. If the effect of the new model deteriorates significantly, it is necessary to retrain. If the new model is better, the current model is replaced for the next cycle of incremental training, and feedback from new users is collected. If there is an error, the model is updated instantly with the sample incrementally, and the distribution changes of the training data are monitored. If concept drift occurs, the incremental learning mechanism needs to be adjusted. Through this periodic incremental learning process, the model can dynamically keep pace with the latest data to achieve effective gradual updates.

[0090] The method for predicting the risk of electricity bill recovery provided in the present application first obtains the first electricity consumption curve of the target user to be predicted within a preset time range, and extracts the electricity consumption characteristics corresponding to the first electricity consumption curve, then obtains the first electricity bill payment data of the target user within the preset time range, and extracts the electricity bill payment characteristics corresponding to the first electricity bill payment data, and then determines the electricity bill behavior characteristics of the target user according to the electricity consumption characteristics and the electricity bill payment characteristics, then obtains the first basic information characteristics of the target user, and inputs the electricity bill behavior characteristics of the target user and the first basic information characteristics into the risk prediction model obtained through training to obtain the electricity bill recovery risk corresponding to the target user, and performs electricity bill recovery risk prediction through the risk prediction model obtained through training, thereby further improving the accuracy and efficiency of electricity bill recovery risk prediction, and at the same time, through further processing the electricity consumption curve and the first electricity bill payment data of the target user, the electricity bill behavior characteristics of the target user are obtained, and compared with directly inputting the electricity consumption curve and the first electricity bill payment data into the risk prediction model, the characteristics of the electricity bill payment behavior pattern are obtained through the hierarchical kernel PCA and hierarchical clustering methods, which reflects the user's long-term electricity bill payment behavior, and further improves the accuracy of the output result of the risk prediction model.

[0091] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a device for predicting electricity fee recovery risks.

[0092] refer to Figure 3 The device for predicting the risk of electricity fee recovery comprises:

[0093] The first acquisition module 201 acquires a first power consumption curve of a target user to be predicted within a preset time range, and extracts power consumption features corresponding to the first power consumption curve;

[0094] The second acquisition module 202 acquires the electricity payment data of the target user to be predicted within a preset time range, and extracts the electricity payment features corresponding to the electricity payment data;

[0095] A determination module 203 determines the electricity fee behavior characteristics of the target user based on the electricity usage characteristics and the electricity fee payment characteristics;

[0096] The third acquisition module 204 acquires the first basic information feature of the target user;

[0097] The prediction module 205 inputs the electricity charge behavior characteristics of the target user and the first basic information characteristics into the risk prediction model obtained through training to obtain the electricity charge recovery risk corresponding to the target user.

[0098] In some embodiments, the first acquisition module is specifically used to:

[0099] Extracting key indicator values ​​of preset indicators corresponding to the first power consumption curve;

[0100] In response to determining that the key indicator value belongs to a preset value range, a key nonlinear feature corresponding to the preset value range is added to the first power consumption curve.

[0101] In some embodiments, the determination module specifically includes:

[0102] A segmentation unit, dividing the preset time range into a plurality of preset time periods;

[0103] An analysis unit, which determines the principal component characteristics corresponding to the electricity usage characteristics and the electricity fee payment characteristics in each preset time period by using a kernel principal component analysis technique;

[0104] A first determining unit, which determines the electricity fee payment mode corresponding to the target user in each preset time period based on the principal component characteristics of each preset time period;

[0105] The second determining unit determines the electricity fee behavior characteristics of the target user based on the electricity fee payment mode corresponding to the target user in each preset time period.

[0106] In some embodiments, the first determining unit is specifically configured to:

[0107] Obtain target features corresponding to each preset payment mode;

[0108] For each preset time period, the distance between the principal component feature and the target feature corresponding to each preset payment mode is determined, and the preset payment mode corresponding to the target feature with the closest distance is determined as the electricity fee payment mode of the target user in the preset time period. In some embodiments, the preset electricity fee payment modes include: advance payment mode, on-time payment mode, intermittent payment mode, delayed payment mode, default payment mode, and overdue power outage mode.

[0109] In some embodiments, the device further comprises a model training module for:

[0110] The parameters of the risk prediction model are adjusted by Monte Carlo tree search technology.

[0111] In some embodiments, the apparatus further comprises an incremental training module for:

[0112] In response to determining that the interval between a first time point corresponding to the end point of the preset time range and a second time point at which the risk prediction model completes training is greater than a preset time period, obtaining a second electricity usage curve, second electricity fee payment data, and second basic information characteristics of the target user from the second time point to the first time point;

[0113] The risk prediction model trained at the second time point is incrementally trained based on the electricity charge behavior characteristics obtained from the second electricity consumption curve and the second electricity charge payment data and the second basic information characteristics.

[0114] For the convenience of description, the above device is described in terms of functions divided into various modules. Of course, when implementing the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0115] The device for predicting the risk of electricity fee recovery in the above-mentioned embodiment is used to implement the corresponding method for predicting the risk of electricity fee recovery in any of the above-mentioned embodiments, and has the beneficial effects of the corresponding method for predicting the risk of electricity fee recovery, which will not be repeated here.

[0116] Based on the same inventive concept, corresponding to any of the above-mentioned embodiments and methods, the present disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for predicting the risk of electricity fee recovery described in any of the above embodiments is implemented.

[0117] It should be noted that the method of the embodiment of the present application can be performed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present application, and the multiple devices will interact with each other to complete the described method.

[0118] It should be noted that the above describes some embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the above embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0119] Figure 4 A more specific schematic diagram of the hardware structure of an electronic device provided in this embodiment is shown. The device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040 and a bus 1050. The processor 1010, the memory 1020, the input / output interface 1030 and the communication interface 1040 are connected to each other in communication within the device through the bus 1050.

[0120] The processor 1010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0121] The memory 1020 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory 1020 and are called and executed by the processor 1010.

[0122] The input / output interface 1030 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0123] The communication interface 1040 is used to connect a communication module (not shown) to realize communication interaction between the device and the power equipment. The communication module can realize communication through a wired method (such as USB, network cable, etc.) or a wireless method (such as mobile network, WIFI, Bluetooth, etc.).

[0124] The bus 1050 includes a path that transmits information between the various components of the device (eg, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).

[0125] It should be noted that, although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiments of the present specification, and does not necessarily include all the components shown in the figure.

[0126] The electronic device of the above embodiment is used to implement the corresponding electricity fee recovery risk prediction method in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.

[0127] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the present disclosure also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the method for predicting the electricity fee recovery risk as described in any of the above embodiments.

[0128] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0129] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the method for predicting the risk of electricity fee recovery as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0130] Based on the same inventive concept, corresponding to any of the above-mentioned embodiments, the present application also provides a computer program product, which includes a computer program. In some embodiments, the computer program instructions can be executed by one or more processors of a computer so that the computer and / or the processor execute the method for predicting the risk of electricity fee recovery described in the above-mentioned embodiments. Corresponding to the execution subject corresponding to each step in each embodiment of the method for predicting the risk of electricity fee recovery, the processor that executes the corresponding step may belong to the corresponding execution subject.

[0131] To simplify the description and discussion, and in order not to make the embodiments of the present application difficult to understand, the known power supply / ground connection with the integrated circuit (IC) chip and other components may or may not be shown in the provided drawings. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present application difficult to understand, and this also takes into account the following fact, that is, the details of the implementation of these block diagram devices are highly dependent on the platform to be implemented in the embodiments of the present application (that is, these details should be fully within the scope of understanding of those skilled in the art). In the case of elaborating specific details (e.g., circuit) to describe exemplary embodiments of the present application, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details or when these specific details are changed. Therefore, these descriptions should be considered to be illustrative rather than restrictive.

[0132] Although the present application has been described in conjunction with specific embodiments of the present application, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0133] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present application (including the claims) is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the present application as described above, which are not provided in detail for the sake of simplicity.

[0134] The embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present application should be included in the scope of protection of the present application.

Claims

1. A method for predicting electricity fee recovery risk, characterized in that: include: Obtaining a first power consumption curve of a target user to be predicted within a preset time range, and extracting power consumption characteristics corresponding to the first power consumption curve; Acquire first electricity fee payment data of the target user within the preset time range, and extract electricity fee payment features corresponding to the first electricity fee payment data; Determining the electricity fee behavior characteristics of the target user based on the electricity usage characteristics and the electricity fee payment characteristics; Acquire a first basic information feature of the target user; The electricity charge behavior characteristics of the target user and the first basic information characteristics are input into the risk prediction model obtained through training to obtain the electricity charge recovery risk corresponding to the target user.

2. The method according to claim 1, characterized in that: Extracting the electricity consumption characteristics corresponding to the first electricity consumption curve specifically includes: Extracting key indicator values ​​of preset indicators corresponding to the first power consumption curve; In response to determining that the key indicator value belongs to a preset value range, a key nonlinear feature corresponding to the preset value range is added to the first power consumption curve.

3. The method according to claim 1, characterized in that Determining the target user's electricity fee behavior characteristics based on the electricity usage characteristics and the electricity fee payment characteristics specifically includes: Dividing the preset time range into a plurality of preset time periods; Determine the principal component characteristics corresponding to the electricity consumption characteristics and the electricity fee payment characteristics in each preset time period by using the kernel principal component analysis technology; Determine the electricity fee payment mode corresponding to each preset time period of the target user based on the principal component characteristics of each preset time period; The electricity fee behavior characteristics of the target user are determined based on the electricity fee payment mode corresponding to the target user in each preset time period.

4. The method according to claim 3, characterized in that: Determining the electricity fee payment mode corresponding to each preset time period by the target user based on the principal component characteristics of each preset time period specifically includes: Obtain target features corresponding to each preset payment mode; For each preset time period, the distance between the principal component feature and the target features corresponding to each preset payment mode is determined, and the preset payment mode corresponding to the target feature with the closest distance is determined as the electricity fee payment mode of the target user in the preset time period.

5. The method according to claim 4, characterized in that The preset electricity fee payment modes include: advance payment mode, on-time payment mode, intermittent payment mode, delayed payment mode, default payment mode, and overdue power outage mode.

6. The method according to claim 1, characterized in that The process of training the risk prediction model includes: The parameters of the risk prediction model are adjusted by Monte Carlo tree search technology.

7. The method according to claim 1, characterized in that Before inputting the target user's electricity charge behavior characteristics and the first basic information characteristics into the risk prediction model obtained through training, the method further includes: In response to determining that the interval between a first time point corresponding to the end point of the preset time range and a second time point at which the risk prediction model completes training is greater than a preset time period, obtaining a second electricity usage curve, second electricity fee payment data, and second basic information characteristics of the target user from the second time point to the first time point; The risk prediction model trained at the second time point is incrementally trained based on the second electricity usage curve, the second electricity bill payment data, and the second basic information features.

8. A device for predicting electricity fee recovery risk, characterized in that: include: A first acquisition module, acquiring a first power consumption curve of a target user to be predicted within a preset time range, and extracting power consumption features corresponding to the first power consumption curve; A second acquisition module acquires first electricity fee payment data of a target user to be predicted within a preset time range, and extracts electricity fee payment features corresponding to the first electricity fee payment data; A determination module, which determines the electricity fee behavior characteristics of the target user based on the electricity usage characteristics and the electricity fee payment characteristics; A third acquisition module is used to acquire the first basic information feature of the target user; The prediction module inputs the electricity charge behavior characteristics of the target user and the first basic information characteristics into the risk prediction model obtained through training to obtain the electricity charge recovery risk corresponding to the target user.

9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method according to any one of claims 1 to 7.

Images