Method and system for identifying risk level of power grid payment user
By acquiring and processing the initial account information of grid payment users, and using decision tree models and clustering algorithms to identify risk levels, the problem of fragmented anti-fraud and anti-money laundering work in the electricity market has been solved, data interoperability and real-time risk identification have been achieved, and the grid's prevention and control capabilities have been improved.
Patent Information
- Application Number
- CN202511559972.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-29
- Publication Date
- 2026-03-03
AI Technical Summary
In the electricity market, anti-fraud and anti-money laundering efforts are scattered across different systems and processes, resulting in low efficiency in collaborative processing. Furthermore, it is difficult to exchange abnormal account transaction information from external institutions with internal grid data in real time, leading to untimely and incomplete risk identification.
By acquiring initial payment account information, performing data preprocessing and feature extraction, using the optimal decision tree model to identify user types, and combining clustering algorithms to determine risk levels, a system for identifying the risk levels of power grid payment users is established, enabling data association and risk assessment.
It has improved the overall effectiveness of anti-fraud and anti-money laundering efforts, increased the efficiency of handling risk events, safeguarded the financial security of power grid companies and the interests of users, and achieved comprehensive risk identification for electricity bill payment services.
Smart Images
Figure CN121599668A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power risk prevention and control, and more specifically, to a method and system for identifying the risk level of power grid payment users. Background Technology
[0002] With the continuous development of the electricity market and the increasing frequency of electricity transactions, more and more power companies have implemented anti-fraud and anti-money laundering risk control measures for electricity bills. However, these measures are often fragmented, with anti-fraud and anti-money laundering work scattered across different systems and processes. This results in each department operating independently in its business processes. Faced with complex payment scenarios, the efficiency of collaborative handling involving multiple parties is low, prone to loopholes and duplication of work, and makes it difficult to form a strong and rapid joint force to prevent and control fraud and money laundering. On the other hand, abnormal account transaction information from external institutions such as anti-fraud centers and financial institutions is difficult to exchange and integrate with customer payment data within the power grid in real time, forming data silos and leading to untimely and incomplete risk identification. Summary of the Invention
[0003] To address the technical problems in existing technologies, such as the low efficiency of collaborative processing due to the fragmentation of anti-fraud and anti-money laundering work in different systems and processes in the electricity market, and the untimely and incomplete risk identification caused by the difficulty in sharing and integrating abnormal account transaction information with external institutions, this invention provides a method and system for identifying the risk level of electricity grid payment users.
[0004] According to one aspect of the present invention, the present invention provides a method for identifying the risk level of a power grid payment user, comprising:
[0005] Obtain the initial payment account information of the user to be identified;
[0006] The initial payment account information is preprocessed to generate valid payment account information;
[0007] Feature extraction is performed on the valid payment account information to generate payment feature data;
[0008] The payment feature data is input into the constructed optimal decision tree model to determine the identification result of the user to be identified, and the probability value corresponding to the identification result. The identification result includes abnormal payment users and normal payment users.
[0009] Based on the established probability-risk score mapping table, the risk score of the user to be identified is determined according to the probability value.
[0010] Based on the identification results and risk scores of the users to be identified, a clustering algorithm is used to cluster the users to be identified, and the risk level of the users to be identified is determined based on the clustering results.
[0011] According to another aspect of the present invention, the present invention provides a system for identifying the risk level of electricity grid payment users, the system comprising:
[0012] The data acquisition module is used to acquire the initial payment account information of the user to be identified;
[0013] The data preprocessing module is used to preprocess the initial payment account information to generate valid payment account information;
[0014] The feature extraction module is used to extract features from the valid payment account information and generate payment feature data;
[0015] The user identification module is used to input the payment feature data into the constructed optimal decision tree model, determine the identification result of the user to be identified, and the probability value corresponding to the identification result, wherein the identification result includes abnormal payment users and normal payment users;
[0016] The risk scoring module is used to determine the risk score of the user to be identified based on the probability value according to the set probability-risk score mapping table.
[0017] The risk level module is used to cluster users based on their identification results and risk scores, and then determine the risk level of each user based on the clustering results.
[0018] According to another aspect of the present invention, a computer-readable storage medium is provided, the storage medium storing a computer program for performing the methods described in any of the above aspects of the present invention.
[0019] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the instructions to implement the method described in any of the preceding aspects of the present invention.
[0020] The present invention discloses a method and system for identifying the risk level of power grid payment users. The method includes: acquiring initial payment account information of the user to be identified; performing data preprocessing on the initial payment account information to generate valid payment account information; extracting features from the valid payment account information to generate payment feature data; inputting the payment feature data into a constructed optimal decision tree model to determine the identification result of the user to be identified and the probability value corresponding to the identification result; determining the risk score of the user to be identified based on the probability value according to a set probability-risk score mapping table; and clustering the user to be identified using a clustering algorithm based on the identification result and risk score, and determining the risk level of the user to be identified based on the clustering result. The proposed method and system break down data barriers between internal power systems and external relevant institutions. It acquires external risk customer information and uses data correlation to identify abnormal payment users in power grid payment accounts. Then, a decision tree model is built using a classification algorithm to predict abnormal users among new electricity payment users. Finally, a clustering algorithm identifies the user's risk level. This achieves information sharing among different systems and institutions, forming a powerful joint prevention and control system. Furthermore, it not only improves the overall effectiveness of anti-fraud and anti-money laundering work but also comprehensively enhances the anti-fraud and anti-money laundering identification capabilities of the power grid's electricity payment business, improves the efficiency of risk event handling, and effectively safeguards the financial security of power grid companies and the interests of users. Attached Figure Description
[0021] Exemplary embodiments of the present invention can be more fully understood by referring to the following figures:
[0022] Figure 1 A flowchart illustrating a method for identifying the risk level of a power grid payment user according to a preferred embodiment of the present invention;
[0023] Figure 2 This is a schematic diagram of the structure of a system for identifying the risk level of electricity grid payment users according to a preferred embodiment of the present invention;
[0024] Figure 3 A hierarchical distributed architecture diagram of a system for identifying the risk level of electricity grid payment users according to a preferred embodiment of the present invention;
[0025] Figure 4 This is a schematic diagram of the structure of an electronic device according to a preferred embodiment of the present invention. Detailed Implementation
[0026] Exemplary embodiments of the invention will now be described with reference to the accompanying drawings. However, the invention may be embodied in many different forms and is not limited to the embodiments described herein. These embodiments are provided to fully and completely disclose the invention and to fully convey its scope to those skilled in the art. The terminology used in the exemplary embodiments illustrated in the drawings is not intended to limit the invention. In the drawings, the same units / elements are referred to by the same reference numerals.
[0027] Unless otherwise stated, the terms used herein (including technical terms) have their common meaning as understood by one of ordinary skill in the art. Furthermore, it is understood that terms defined in commonly used dictionaries should be understood to have a meaning consistent with the context of their relevant field, and not to be interpreted as having an idealized or overly formal meaning.
[0028] Exemplary methods
[0029] Figure 1 This is a flowchart illustrating a method for identifying the risk level of electricity grid payment users according to a preferred embodiment of the present invention. Figure 1 As shown, the method for identifying the risk level of grid payment users in this preferred embodiment begins with step 101.
[0030] In step 101, the initial payment account information of the user to be identified is obtained.
[0031] In this preferred embodiment, in order to identify the risk level of grid payment users, before obtaining the initial payment users of the users to be identified, it is necessary to establish a decision tree model for identifying user types based on historical payment account information that can distinguish between abnormal payment users and normal payment users.
[0032] Preferably, before obtaining the initial payment account information of the user to be identified, the method further includes establishing an optimal decision tree model for identifying the type of power grid payment user, wherein:
[0033] In step 1101, historical payment account information within the power grid and historical risk customer information from third parties within the same time period are obtained. The information items of the historical payment account information include user name, electricity account number, payment area, payment amount, payment time and payment method. The information items of the risk customer information include user name, transaction time, transaction amount, transaction area and transaction method.
[0034] In this preferred embodiment, data access from multiple data sources is achieved by establishing connections with various systems within the power grid and third-party institutions. For example, historical payment account information from the past year can be obtained from the financial and marketing domains within the power grid. Correspondingly, blacklist information for the same period can be obtained from the anti-fraud center, and risky user information for the same period can be obtained from financial institutions, which are then used as historical risky customer information for third parties. Once the optimal decision tree model is established, when it is necessary to use the optimal decision tree model to identify new power grid payment account information, the latest power grid payment account information can be directly obtained from the data access interface within the power grid as the initial payment account information for the user to be identified.
[0035] In step 1102, the historical payment account information and the risk customer information are preprocessed to obtain historical valid account information and valid risk customer information.
[0036] In this preferred embodiment, after obtaining the historical payment account information and the risk customer information, the data is cleaned to remove duplicate, erroneous, and incomplete data records, ensuring the accuracy and integrity of the data. Furthermore, the cleaned data is converted to ensure a consistent format; for example, the transaction times in the anti-fraud center's blacklist user data and the transaction times in the bank's risk customer information data are converted from timestamp format to a character format consistent with the payment times in the Southern Power Grid database, facilitating subsequent data processing. After data cleaning and conversion, the historical payment account information and the risk customer information generate historical valid account information and valid risk customer information.
[0037] In step 1103, the valid risk customer information is associated with the corresponding information items in the historical valid account information according to the set association rules to determine the historical valid account information of power grid abnormal payment users and normal payment users.
[0038] Preferably, the valid risk customer information is associated with the corresponding information items in the historical valid account information according to the set association rules to determine the historical valid account information of power grid abnormal payment users and normal payment users. The association rules are to identify the historical valid account information of power grid abnormal payment users as the historical valid account information of power grid abnormal payment users if there are at least two information items that match the valid risk customer information including the user name.
[0039] In this preferred embodiment, since the users in the valid risk customer information have all been confirmed as fraudulent users, if the user name is the same as the user name in the historical valid account information within the power grid, and one of the other information items is the same, it can be basically determined that the user is an abnormal payment user of the power grid. However, in order to improve the accuracy of identification, the number of information items for identifying abnormal payment users can be increased in addition to the user name.
[0040] In step 1104, feature extraction is performed on the historical valid account information to generate historical payment feature data.
[0041] Preferably, feature extraction is performed on the historical valid account information to generate historical payment feature data, including:
[0042] Generate user payment frequency behavior characteristics based on the user name and payment time in the historical valid account information;
[0043] Generate user payment amount behavior characteristics based on the user name and payment amount in the historical valid account information;
[0044] Generate user payment channel count behavior characteristics based on the username and payment method in the historical valid account information;
[0045] Generate user payment region count behavior characteristics based on the user name and payment region in the historical valid account information;
[0046] Generate the number of user-associated account numbers and behavioral characteristics based on the user name and electricity usage code in the historical valid account information.
[0047] In this preferred embodiment, in addition to identifying normal and abnormal payment users from historical valid account information, it is also necessary to extract their payment characteristics. These characteristics are then used to build a binary classification decision tree model for user type identification. This embodiment extracts features based on pairwise combinations of user names and other feature items. However, in practical applications, other methods can also be used, and the examples in this invention do not constitute a limitation.
[0048] In step 1105, the historical payment characteristic data of users with normal payment records are used as positive samples, and the historical payment characteristic data of users with abnormal payment records are used as negative samples to generate a sample set, and the sample set is divided into a training set and a validation set.
[0049] In step 1106, based on the historical payment feature data in the training set, the initial decision tree model constructed using the CART classification tree algorithm is trained to generate a candidate decision tree model.
[0050] The first step in building a decision tree is to select payment features for data splitting. For each node, the decision tree algorithm evaluates all possible features and selects the best feature for splitting based on the chosen payment feature evaluation criteria. Feature evaluation criteria depend on the algorithm used and mainly include information gain, information gain ratio, and Gini coefficient. This invention uses these three criteria to build decision tree models and tests them, confirming that the CART classification tree algorithm using the Gini coefficient is the optimal algorithm. During the training of the decision tree model, necessary information is saved to generate a model file. The model file mainly contains node information, feature list, predicted values, and topology, such as the child nodes of each node and the branch logic corresponding to each splitting condition.
[0051] In step 1107, the historical payment feature data in the verification set is input into the candidate decision tree model to determine the identification result.
[0052] In this preferred embodiment, the model file saved during training can be parsed to reconstruct the decision tree model generated during training. Payment feature data from the validation set is input into the candidate decision tree model, and traversal begins from the root node. At each internal node, the data is judged according to the node's splitting condition. For example, if the splitting condition is "payment frequency > 10 within a time period," the value of the payment frequency behavior feature in the payment feature data is compared with the value of 10. Based on the judgment result, the corresponding branch (left branch or right branch) is selected to proceed to the next node. This process continues until a leaf node is reached, thereby obtaining the user identification result and the corresponding probability value, where the probability value ranges from [0,1].
[0053] In step 1108, the identification results are verified based on preset test indicators to generate verification results.
[0054] Preferably, the recognition result is verified based on preset test metrics to generate a verification result, wherein the test metrics include accuracy A, precision P, recall R, and F1-score, and their expressions are as follows:
[0055] A = (TP + TN) / (TP + FP + FN + TN)
[0056] P = TP / (TP + FP)
[0057] R = TP / (TP + FN)
[0058] F1 score = 2PR / (P+R)
[0059] In the formula, variables TP and FP represent users whose identification results are both normal payment users, but in the validation set, the former is a normal payment user and the latter is an abnormal payment user. TN and FN represent users whose identification results are both abnormal payment users, but in the validation set, the former is an abnormal payment user and the latter is a normal payment user.
[0060] In this preferred embodiment, precision A represents the proportion of the number of correctly identified samples to the total number of samples in the validation set, accuracy P represents the proportion of the number of correctly identified positive samples to the proportion of all samples in the validation set that are identified as normal paying users, recall R represents the proportion of the number of correctly identified positive samples to the proportion of all positive samples in the validation set, and F1-scor is used to harmonic average precision and recall, which is especially useful when the classes are imbalanced.
[0061] In step 1109, when the verification result meets the preset model testing rules, the candidate decision tree model is determined to be the optimal decision tree model for identifying the risk level of grid payment users. The model testing rules refer to the fact that the values of all test indicators in the verification result are within their respective value ranges.
[0062] In this preferred embodiment, based on the selected test index, the result value of the test index is calculated according to the identification result, and the result value is compared with the model test rule. For example, if the model test rule stipulates that A is not less than 0.98, then when the result value of A is 0.99, it is considered that the result value of test index A meets the model test rule. If the result values of other test indices also meet the judgment conditions in the corresponding model test rule, then the candidate decision tree model can be identified as the optimal decision tree model.
[0063] In step 102, the initial payment account information is preprocessed to generate valid payment account information.
[0064] In step 103, feature extraction is performed on the valid payment account information to generate payment feature data.
[0065] In this preferred embodiment, the data preprocessing of the initial payment account information is the same as the steps of cleaning and transforming the historical payment account information, and the feature extraction of the valid payment account information is also the same as the steps of feature extraction of the historical valid account information, which will not be described again here.
[0066] In step 104, the payment feature data is input into the constructed optimal decision tree model to determine the identification result of the user to be identified, and the probability value corresponding to the identification result. The identification result includes abnormal payment users and normal payment users.
[0067] In step 105, based on the set probability-risk score mapping table, the risk score of the user to be identified is determined according to the probability value.
[0068] Preferably, the risk score of the user to be identified is determined based on the probability value according to the set probability-risk score mapping table, wherein the probability-risk score mapping table refers to mapping the interval [0,1] where the probability value output by the optimal decision tree model is located to the risk score interval [0,100].
[0069] In step 106, based on the identification results and risk scores of the users to be identified, a clustering algorithm is used to cluster the users to be identified, and the risk level of the users to be identified is determined based on the clustering results.
[0070] Preferably, the step of clustering the users to be identified using a clustering algorithm based on the identification results and risk scores, and determining the risk level of the users to be identified based on the clustering results, includes:
[0071] Based on business needs, the risk levels of power grid users are classified into Category K.
[0072] Based on the identification results and risk scores of the users to be identified, the K-means clustering algorithm is used to cluster the users to be identified and obtain the clustering results. The number of clusters is K, and the clustering results include the cluster to which the risk score of each user to be identified belongs and the cluster center of each cluster.
[0073] The users to be identified with the corresponding risk level can be obtained by sorting the cluster centers of each cluster in descending order of their mean values.
[0074] In this preferred embodiment, based on the identification results and risk scores of the users to be identified, the K-means clustering algorithm is used to cluster the users to be identified, dividing the users' risk levels into 5 categories, i.e., K=5, with risk levels being high risk, medium-high risk, medium risk, medium-low risk, and low risk. The specific implementation method is as follows:
[0075] Initialize centroids by randomly selecting the identification results and risk scores of 5 users to be identified as initial cluster centers. These centroids will serve as representatives of each cluster.
[0076] Data points are assigned by iterating through all the identification results and risk scores of the users to be identified, and each data point is assigned to the nearest cluster center, typically using Euclidean distance as the distance metric.
[0077]
[0078] In the formula, c is the coordinate of the cluster center, and x is the coordinate of the data point.
[0079] Update cluster centers by recalculating the cluster centers for all data points in each cluster. The new cluster centers are the mean of all data points within that cluster.
[0080]
[0081] In the formula, N is the number of data points within the cluster, and x i These are data points within a cluster.
[0082] Repeat the steps of assigning data points and updating cluster centers, continue to assign data points to new cluster centers, and update cluster centers until the cluster centers no longer change (convergence), or reach the preset number of iterations.
[0083] Output results: Output the final clustering results, including the cluster to which each data point belongs and the cluster center of each cluster. Finally, the user groups with different risk levels are obtained by taking the mean of the cluster centers in descending order.
[0084] After determining the risk level of the user to be identified, if the user is at a high risk level, relevant information of the high-risk user, such as user information, risk details and payment information, can be transmitted to the business front desk and relevant business personnel can be reminded to verify and process it. This achieves risk warning, makes it convenient for staff to keep abreast of the user's risk status in real time, and reminds relevant personnel to pay special attention to and investigate high-risk users to strengthen risk prevention and control.
[0085] The method for identifying the risk level of power grid payment users described in this preferred embodiment integrates multi-channel data and an intelligent risk assessment model, enabling real-time monitoring and identification of fraud and money laundering risks in electricity bill payment transactions. Furthermore, by combining decision trees and K-Means algorithms, the advantages of both algorithms can be fully utilized, overcoming the limitations of a single algorithm and further improving the accuracy of risk assessment for payment users. Compared with traditional methods, the risk identification time is shortened, the accuracy rate is improved, the occurrence of risk events is effectively reduced, and the financial security of power grid companies and the interests of users are protected.
[0086] Exemplary System
[0087] Figure 2 This is a schematic diagram of a system for identifying the risk level of electricity grid payment users according to a preferred embodiment of the present invention. Figure 2 As shown, the system 200 for identifying the risk level of electricity grid payment users according to this preferred embodiment includes:
[0088] Data acquisition module 201 is used to acquire the initial payment account information of the user to be identified;
[0089] Data preprocessing module 202 is used to preprocess the initial payment account information to generate valid payment account information;
[0090] Feature extraction module 203 is used to extract features from the valid payment account information and generate payment feature data;
[0091] User identification module 204 is used to input the payment feature data into the constructed optimal decision tree model to determine the identification result of the user to be identified and the probability value corresponding to the identification result, wherein the identification result includes abnormal payment users and normal payment users;
[0092] Risk scoring module 205 is used to determine the risk score of the user to be identified based on the probability value according to the set probability-risk score mapping table;
[0093] The risk level module 206 is used to cluster the users to be identified based on their risk scores using a clustering algorithm, and to determine the risk level of the users to be identified based on the clustering results.
[0094] Preferably, the system further includes a module construction module for establishing an optimal decision tree model for identifying the type of electricity grid payment user, wherein:
[0095] The system obtains historical payment account information within the power grid and historical risk customer information from third parties for the same time period. The information items of the historical payment account information include user name, electricity account number, payment area, payment amount, payment time and payment method. The information items of the risk customer information include user name, transaction time, transaction amount, transaction area and transaction method.
[0096] The historical payment account information and the risk customer information are preprocessed to obtain historical valid account information and valid risk customer information;
[0097] The valid risk customer information is associated with the corresponding information items in the historical valid account information according to the set association rules to determine the historical valid account information of power grid abnormal payment users and normal payment users;
[0098] Feature extraction is performed on the historical valid account information to generate historical payment feature data;
[0099] The historical payment characteristic data of users with normal payment records are used as positive samples, and the historical payment characteristic data of users with abnormal payment records are used as negative samples to generate a sample set, which is then divided into a training set and a validation set.
[0100] Based on the historical payment feature data in the training set, the initial decision tree model constructed using the CART classification tree algorithm is trained to generate candidate decision tree models.
[0101] Input the historical payment feature data from the validation set into the candidate decision tree model to determine the identification result;
[0102] The identification results are verified based on preset test indicators to generate verification results;
[0103] When the verification results meet the preset model testing rules, the candidate decision tree model is determined to be the optimal decision tree model for identifying the risk level of grid payment users. The model testing rules refer to the fact that the values of all test indicators in the verification results are within their respective value ranges.
[0104] Preferably, the model building module associates the valid risk customer information with the corresponding information items in the historical valid account information according to the set association rules to determine the historical valid account information of power grid abnormal payment users and normal payment users. The association rules identify historical valid account information that matches at least two information items including the user name in the valid risk customer information as the historical valid account information of power grid abnormal payment users.
[0105] Preferably, the model building module extracts features from the historical valid account information to generate historical payment feature data, including:
[0106] Generate user payment frequency behavior characteristics based on the user name and payment time in the historical valid account information;
[0107] Generate user payment amount behavior characteristics based on the user name and payment amount in the historical valid account information;
[0108] Generate user payment channel count behavior characteristics based on the username and payment method in the historical valid account information;
[0109] Generate user payment region count behavior characteristics based on the user name and payment region in the historical valid account information;
[0110] Generate the number of user-associated account numbers and behavioral characteristics based on the user name and electricity usage code in the historical valid account information.
[0111] Preferably, the model building module verifies the recognition results based on preset test metrics and generates verification results, wherein the test metrics include accuracy A, precision P, recall R, and F1-score, and their expressions are as follows:
[0112] A = (TP + TN) / (TP + FP + FN + TN)
[0113] P = TP / (TP + FP)
[0114] R = TP / (TP + FN)
[0115] F1 score = 2PR / (P+R)
[0116] In the formula, variables TP and FP represent users whose identification results are both normal payment users, but in the validation set, the former is a normal payment user and the latter is an abnormal payment user. TN and FN represent users whose identification results are both abnormal payment users, but in the validation set, the former is an abnormal payment user and the latter is a normal payment user.
[0117] Preferably, the risk scoring module 205 determines the risk score of the user to be identified based on the probability value according to the set probability-risk score mapping table, wherein the probability-risk score mapping table refers to mapping the interval [0,1] where the probability value output by the optimal decision tree model is located to the risk score interval [0,100].
[0118] Preferably, the risk level module 206 clusters the users to be identified using a clustering algorithm based on the identification results and risk scores of the users to be identified, and determines the risk level of the users to be identified based on the clustering results, including:
[0119] Based on business needs, the risk levels of power grid users are classified into Category K.
[0120] Based on the identification results and risk scores of the users to be identified, the K-means clustering algorithm is used to cluster the users to be identified and obtain the clustering results. The number of clusters is K, and the clustering results include the cluster to which the risk score of each user to be identified belongs and the cluster center of each cluster.
[0121] The users to be identified with the corresponding risk level can be obtained by sorting the cluster centers of each cluster in descending order of their mean values.
[0122] After determining the risk level of the user to be identified, the system described in this preferred embodiment can also generate risk warning information, and display the relevant warning information and the relevant payment data of high-risk users on the user interaction platform, as well as perform risk handling, data query and other services through the user interaction platform.
[0123] Figure 3 This is a hierarchical distributed architecture diagram of a system for identifying the risk level of electricity grid payment users according to a preferred embodiment of the present invention. Figure 3As shown, the data access layer is equivalent to a data acquisition module, used to obtain payment account information and user payment information from the financial and marketing domains within the power grid, respectively, and to obtain anti-fraud blacklist information and risk customer information from third parties and financial institutions. The data processing layer is equivalent to a data preprocessing module and a feature extraction module, used to clean and process the data to generate payment feature data. The risk assessment layer is equivalent to a user identification module, a risk scoring module, and a risk level module, used to assess the risk of user payment behavior based on the established optimal decision tree model and clustering algorithm to determine the user's risk level. Furthermore, after determining the user's risk level, a risk warning can be issued based on the risk level, and the risk warning information can be sent to the front-end user interaction layer.
[0124] The system for identifying the risk level of grid payment users described in this preferred embodiment and the method for identifying the risk level of grid payment users use the same steps of establishing a decision tree model and clustering algorithm to identify the risk level of grid payment users, and achieve the same technical effect, so they will not be described again here.
[0125] Exemplary electronic devices
[0126] Figure 4 This is a schematic diagram of the structure of an electronic device according to a preferred embodiment of the present invention. Figure 4 As shown, the electronic device includes one or more processors 401 and memory 402.
[0127] The processor 401 may be a central processing unit (CPU) or other form of processing unit with data processing and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.
[0128] The memory 402 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 401 may execute the program instructions to implement the methods for identifying the risk level of grid payment users in the various embodiments disclosed above, and / or other desired functions. In one example, the electronic device may also include an input device 403 and an output device 404, these components being interconnected via a bus system and / or other forms of connection mechanisms (not shown).
[0129] In addition, the input device 403 may also include, for example, a keyboard, a mouse, etc.
[0130] The output device 404 can output various information to the outside. The output device 404 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.
[0131] Of course, for the sake of simplicity, Figure 4 Only some of the components of the electronic device relevant to this disclosure are shown, omitting components such as buses, input / output interfaces, etc. In addition, the electronic device may include any other suitable components depending on the specific application.
[0132] Exemplary computer program products and computer-readable storage media
[0133] In addition to the methods and apparatus described above, embodiments of this disclosure may also be computer program products comprising computer program instructions that, when executed by a processor, cause the processor to perform the steps of the methods for identifying the risk level of grid payment users according to various embodiments of this disclosure as described in the "Exemplary Methods" section of this specification.
[0134] The computer program product can be written in any combination of one or more programming languages to perform the operations of the embodiments of this disclosure. The programming languages include object-oriented programming languages such as Java and C++, as well as conventional procedural programming languages such as C or similar languages. The program code can be executed entirely on a user's computing device, partially on a user's computing device, as a standalone software package, partially on a user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.
[0135] Furthermore, embodiments of this disclosure may also be computer-readable storage media storing computer program instructions thereon, which, when executed by a processor, cause the processor to perform the steps in the methods for identifying the risk level of grid payment users according to various embodiments of this disclosure as described in the "Exemplary Methods" section above.
[0136] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.
[0137] The basic principles of this disclosure have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this disclosure are merely examples and not limitations, and should not be considered as essential features of each embodiment of this disclosure. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the scope of this disclosure to the necessity of employing the aforementioned specific details for implementation.
[0138] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For system embodiments, since they largely correspond to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.
[0139] The block diagrams of devices, apparatuses, devices, and systems disclosed herein are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.
[0140] The apparatus and methods of this disclosure may be implemented in many ways. For example, they may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above-described order of steps for the methods is for illustrative purposes only, and the steps of the methods of this disclosure are not limited to the order specifically described above unless otherwise specifically stated. Furthermore, in some embodiments, this disclosure may also be implemented as a program recorded on a recording medium, the program including machine-readable instructions for implementing the methods according to this disclosure. Thus, this disclosure also covers recording media storing programs for performing the methods according to this disclosure.
[0141] It should also be noted that in the apparatus, devices, and methods of this disclosure, the components or steps are decomposable and / or recombinable. Such decomposition and / or recombination should be considered equivalent to the present disclosure. The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this disclosure. Therefore, this disclosure is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.
[0142] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this disclosure to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.
Claims
1. A method for identifying the risk level of electricity grid payment users, characterized in that, The method includes: Obtain the initial payment account information of the user to be identified; The initial payment account information is preprocessed to generate valid payment account information; Feature extraction is performed on the valid payment account information to generate payment feature data; The payment feature data is input into the constructed optimal decision tree model to determine the identification result of the user to be identified, and the probability value corresponding to the identification result. The identification result includes abnormal payment users and normal payment users. Based on the established probability-risk score mapping table, the risk score of the user to be identified is determined according to the probability value. Based on the identification results and risk scores of the users to be identified, a clustering algorithm is used to cluster the users to be identified, and the risk level of the users to be identified is determined based on the clustering results.
2. The method according to claim 1, characterized in that, Before obtaining the initial payment account information of the user to be identified, the process also includes establishing an optimal decision tree model for identifying the type of power grid payment user, where: The system obtains historical payment account information within the power grid and historical risk customer information from third parties for the same time period. The information items of the historical payment account information include user name, electricity account number, payment area, payment amount, payment time and payment method. The information items of the risk customer information include user name, transaction time, transaction amount, transaction area and transaction method. The historical payment account information and the risk customer information are preprocessed to obtain historical valid account information and valid risk customer information; The valid risk customer information is associated with the corresponding information items in the historical valid account information according to the set association rules to determine the historical valid account information of power grid abnormal payment users and normal payment users; Feature extraction is performed on the historical valid account information to generate historical payment feature data; The historical payment characteristic data of users with normal payment records are used as positive samples, and the historical payment characteristic data of users with abnormal payment records are used as negative samples to generate a sample set, which is then divided into a training set and a validation set. Based on the historical payment feature data in the training set, the initial decision tree model constructed using the CART classification tree algorithm is trained to generate candidate decision tree models. The historical payment feature data in the validation set is input into the candidate decision tree model to determine the identification result; The recognition results are verified based on preset test indicators to generate verification results; When the verification results meet the preset model testing rules, the candidate decision tree model is determined to be the optimal decision tree model for identifying the risk level of grid payment users. The model testing rules refer to the fact that the values of all test indicators in the verification results are within their respective value ranges.
3. The method according to claim 2, characterized in that, The valid risk customer information is associated with the corresponding information items in the historical valid account information according to the set association rules to determine the historical valid account information of power grid abnormal payment users and normal payment users. The association rule is to identify the historical valid account information of power grid abnormal payment users as the historical valid account information of power grid abnormal payment users if there are at least two information items that match the valid risk customer information including the user name.
4. The method according to claim 3, characterized in that, Feature extraction is performed on the historical valid account information to generate historical payment feature data, including: Generate user payment frequency behavior characteristics based on the user name and payment time in the historical valid account information; Generate user payment amount behavior characteristics based on the user name and payment amount in the historical valid account information; Generate user payment channel count behavior characteristics based on the username and payment method in the historical valid account information; Generate user payment region count behavior characteristics based on the user name and payment region in the historical valid account information; Generate the number of user-associated account numbers and behavioral characteristics based on the user name and electricity usage code in the historical valid account information.
5. The method according to claim 2, characterized in that, The recognition results are verified based on preset test metrics to generate verification results. These test metrics include accuracy (A), precision (P), recall (R), and F1-score, expressed as follows: A = (TP + TN) / (TP + FP + FN + TN) P = TP / (TP + FP) R = TP / (TP + FN) F1 score = 2PR / (P+R) In the formula, variables TP and FP represent users whose identification results are both normal payment users, but in the validation set, the former is a normal payment user and the latter is an abnormal payment user. TN and FN represent users whose identification results are both abnormal payment users, but in the validation set, the former is an abnormal payment user and the latter is a normal payment user.
6. The method according to claim 1, characterized in that, The probability-risk score mapping table is set to determine the risk score of the user to be identified based on the probability value. The probability-risk score mapping table refers to mapping the interval [0,1] where the probability value output by the optimal decision tree model is located to the risk score interval [0,100].
7. The method according to claim 1, characterized in that, The process of clustering users based on their identification results and risk scores, and determining their risk levels based on the clustering results, includes: Based on business needs, the risk levels of power grid users are classified into Category K. Based on the identification results and risk scores of the users to be identified, the K-means clustering algorithm is used to cluster the users to be identified and obtain the clustering results. The number of clusters is K, and the clustering results include the cluster to which the risk score of each user to be identified belongs and the cluster center of each cluster. The users to be identified with the corresponding risk level can be obtained by sorting the cluster centers of each cluster in descending order of their mean values.
8. A system for identifying the risk level of electricity grid payment users, characterized in that, The system includes: The data acquisition module is used to acquire the initial payment account information of the user to be identified; The data preprocessing module is used to preprocess the initial payment account information to generate valid payment account information; The feature extraction module is used to extract features from the valid payment account information and generate payment feature data; The user identification module is used to input the payment feature data into the constructed optimal decision tree model, determine the identification result of the user to be identified, and the probability value corresponding to the identification result, wherein the identification result includes abnormal payment users and normal payment users; The risk scoring module is used to determine the risk score of the user to be identified based on the probability value according to the set probability-risk score mapping table. The risk level module is used to cluster users to be identified based on their risk scores using a clustering algorithm, and to determine the risk level of the users to be identified based on the clustering results.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method as described in any one of claims 1-7.
10. An electronic device, characterized in that, include: The computer-readable storage medium as described in claim 9; as well as One or more processors for executing a program in the computer-readable storage medium.