Fraud gang identification method based on knowledge graph

Abstract

The invention relates to the technical field of science and technology finance, and provides a fraud gang identification method based on a knowledge graph, which constructs and optimizes a fraud identification model by continuously acquiring user identity verification and transaction data. According to the method, the model performance is improved by applying the genetic algorithm, and the recognition capability of complex fraud gangs is enhanced. According to the invention, high-risk accounts can be monitored and early warned in real time, suspicious transactions are stopped in time, and corresponding anti-fraud actions are executed according to feedback. According to the method, the problem that complex association and behavior modes are difficult to capture in a traditional method is solved, the accuracy and efficiency of fraud gang recognition are remarkably improved, and powerful guarantee is provided for financial security.

Description

Technical Field

[0001] The present invention relates to the field of science and technology financial technology, and in particular to a method for identifying fraud gangs based on knowledge graphs. Background Art

[0002] At present, the identification of fraud gangs usually relies on traditional data mining technology and machine learning algorithms. These methods are mainly based on the extraction of data features and pattern recognition, such as abnormal detection of transaction behavior and analysis of user behavior. However, these methods have the following limitations: For complex fraud gangs, a single data feature and pattern is often difficult to capture their hidden associations and behavior patterns. There is a lack of modeling of complex relationships between entities, such as the associations and behavior patterns between members of a fraud gang.

[0003] In different financial fields, the behavioral characteristics of fraud gangs vary greatly, and traditional methods are difficult to adapt to the fraud identification needs of different fields. Existing anti-fraud gang identification has the following shortcomings: Traditional methods often only focus on the abnormality of individual behavior and lack the analysis of the association between gangs. The identification ability for complex gang structures is limited, and it is easy to misjudge or miss reports. For the identification of cross-field fraud gangs, there is a lack of comprehensive analysis and identification methods. Summary of the invention

[0004] The purpose of the present invention is to provide a fraud gang identification method based on knowledge graph to solve the problem of lack of comprehensive analysis and identification means for the identification of financial fraud gangs.

[0005] The present invention provides a method for identifying fraud gangs based on knowledge graph, comprising:

[0006] Step S101, the server side continuously obtains the user's identity authentication data in financial transactions and the user's normal transaction data from the application and the website;

[0007] Step S102, constructing a fraud identification model based on historical fraud data and normal transaction data, and optimizing the fraud identification model using a genetic algorithm to obtain an optimized fraud identification model;

[0008] Step S103: continuously obtain the user's identity verification data in financial transactions and the user's normal transaction data from the application and the website and substitute them into the optimized fraud identification model to obtain identification result information;

[0009] Step S104, determining the transaction account information according to the identity verification data in the identification result information, and when the transaction account information corresponding to the identification result information is traded again, generating an early warning message and sending it to the backend control terminal and the user terminal, and stopping the transaction execution corresponding to the transaction account information;

[0010] Step S105, executing a preset anti-fraud execution action according to the information fed back by the backend control terminal and the user terminal.

[0011] Furthermore, in the method for identifying fraud gangs based on knowledge graph provided by the present invention, in step S101, the server side continuously obtains the user's identity authentication data and the user's normal transaction data in financial transactions from the application and the website, including:

[0012] The server side first needs to set up clear data collection interfaces, which should be compatible with the data output formats of applications and websites to ensure smooth data transmission.

[0013] Establish specifications for data collection and clarify which data is necessary, such as user authentication information (such as username, password hash, dynamic token, etc.), transaction information (such as transaction amount, transaction time, counterparty account, etc.), and device information (such as IP address, device ID, operating system, etc.).

[0014] Set the frequency of data collection to ensure that all important data is captured without affecting the performance of applications and websites.

[0015] The server side continuously fetches data from applications and websites through the set interface. This process needs to ensure the integrity and accuracy of the data.

[0016] Perform preliminary verification on the captured data to check whether the data format is correct and whether there are missing values ​​or outliers to ensure data quality.

[0017] The verified data is stored in the server's database for subsequent analysis.

[0018] Preprocess the collected raw data, including data cleaning (removing duplicate, invalid or erroneous data) and data conversion (converting data into a format suitable for analysis).

[0019] Integrate identity verification data and transaction data to establish connections between the data and form a complete data set. For example, match each transaction with its corresponding identity verification information.

[0020] Further quality checks are carried out on the integrated data to ensure its accuracy and consistency, providing a high-quality data foundation for subsequent identification of fraud gangs.

[0021] Furthermore, the fraud group identification method based on knowledge graph provided by the present invention builds a fraud identification model based on historical fraud data and normal transaction data, and optimizes the fraud identification model using a genetic algorithm to obtain an optimized fraud identification model, including:

[0022] The server first organizes and prepares historical fraud data and normal transaction data to ensure the quality and integrity of the data.

[0023] Perform feature engineering to extract meaningful features from historical data, including transaction frequency, transaction amount distribution, transaction time pattern, user behavior pattern, etc.

[0024] The extracted features are selected and transformed to remove redundant and irrelevant information. At the same time, some techniques such as normalization and standardization are used to process the features to make them more suitable for model training.

[0025] Select a suitable machine learning algorithm or model framework to build an initial fraud identification model, such as decision tree, random forest, neural network and other algorithms.

[0026] Use the prepared dataset to train the model and optimize the performance of the model by adjusting model parameters and learning strategies.

[0027] Preliminarily evaluate the performance of the model, use techniques such as cross-validation to ensure that the model is not overfitting or underfitting, and record the performance metrics of the initial model.

[0028] Define the fitness function of the genetic algorithm, which should be able to reflect the performance of the fraud identification model. For example, accuracy, recall rate, F1 score, etc. can be used as evaluation indicators.

[0029] Initialize the population of the genetic algorithm, where each individual represents a set of parameter configurations for the fraud identification model.

[0030] The optimization process of the genetic algorithm, including operations such as selection, crossover, and mutation, is carried out to find the optimal model parameter configuration through continuous iteration.

[0031] After each generation iteration, the fitness of individuals in the population is evaluated, individuals with high fitness are selected to enter the next generation, and individuals with low fitness are eliminated.

[0032] When the preset number of iterations is reached or other stopping conditions are met, the optimization process of the genetic algorithm is stopped and the optimized fraud recognition model parameter configuration is output.

[0033] Furthermore, the fraud group identification method based on knowledge graph provided by the present invention continuously obtains the user's identity authentication data in financial transactions and the user's normal transaction data from the application and website and substitutes them into the optimized fraud identification model to obtain the identification result information, including:

[0034] The server side continuously obtains the latest user identity authentication data and normal transaction data in financial transactions from applications and websites through the previously set interface.

[0035] Preprocess the newly acquired data to ensure that the format and structure of the data are consistent with the data used in model training, including operations such as data cleaning, conversion, and standardization.

[0036] Verify the completeness and validity of the data and remove any incomplete or invalid data records to ensure the quality of data input into the model.

[0037] Substitute the preprocessed data into the fraud identification model previously optimized by genetic algorithm.

[0038] Run the model to process and analyze the data. The model will determine whether there is a risk of fraud for each transaction based on its internal algorithms and logic.

[0039] Get the output of the model, which typically includes a fraud risk score for each transaction, a predicted label (such as fraudulent or not fraudulent), and other relevant information.

[0040] Parse the output of the model and convert it into a format that can be understood by the server and subsequent processing systems.

[0041] The parsed recognition result information is stored in the server's database for subsequent query and analysis.

[0042] As needed, the identification result information is sent to relevant systems or personnel for further processing or verification.

[0043] Furthermore, the fraud gang identification method based on knowledge graph provided by the present invention determines the transaction account information according to the identity authentication data in the identification result information, and when the transaction account information corresponding to the identification result information is traded again, generates an early warning message and sends it to the background control terminal and the user terminal, and stops the transaction execution corresponding to the transaction account information, including:

[0044] The server performs an in-depth analysis of the recognition result information obtained in step S103, paying particular attention to transactions that are marked by the model as high-risk or suspected fraud.

[0045] Extract identity verification data from high-risk transactions and use this data to accurately determine the account information involved in the transaction.

[0046] Establish a monitoring list and add those trading accounts identified as suspected fraud to the list for subsequent monitoring.

[0047] A real-time monitoring mechanism is set up on the server side to continuously monitor all financial transactions, especially the transaction activities of accounts in the monitoring list.

[0048] When an account in the monitoring list attempts to trade again, the server immediately triggers an early warning mechanism.

[0049] Prepare to generate early warning information, which should include trading account number, transaction details, risk warnings, and recommended actions.

[0050] The server sends the warning information to the backend control and user ends in real time to ensure that relevant personnel and users can know and take action immediately.

[0051] At the same time, the server automatically executes transaction blocking operations to suspend or cancel current transactions of accounts in the monitoring list to prevent potential fraudulent activities from being completed.

[0052] Record the detailed information of the warning and transaction blocking, including time, account number, transaction details, etc., for subsequent audit and analysis.

[0053] Furthermore, the fraud group identification method based on knowledge graph provided by the present invention, step S105, executes a preset anti-fraud execution action according to the information fed back by the backend control terminal and the user terminal, including:

[0054] The server side sets up a receiving mechanism for obtaining feedback information from the background control side and the user side in real time.

[0055] When feedback information arrives, the server immediately receives and parses it to ensure that the content and intent of the feedback are accurately understood.

[0056] Based on the results of the analysis, the feedback information is classified into categories such as confirmed fraud, false alarm, need for further investigation, etc.

[0057] According to the classification of the feedback information, the server queries the preset anti-fraud execution action library, which contains a series of pre-defined execution actions for different situations.

[0058] Determine the execution actions that match the current feedback information, which may include freezing the account suspected of fraud, notifying relevant departments for further investigation, lifting restrictions on false positive accounts, etc.

[0059] If the feedback information indicates that further investigation or human intervention is required, the server will trigger the corresponding process, such as transferring the case to a dedicated investigation team.

[0060] The server automatically executes certain anti-fraud actions, such as freezing accounts, sending notifications, etc.

[0061] While executing the action, the server will generate detailed execution records, including execution time, execution action type, account information involved, etc. These records will be stored in the server's log system for subsequent auditing, analysis and improvement of the anti-fraud system.

[0062] Beneficial effects of the present invention:

[0063] Improve the accuracy of fraud group identification: By combining the knowledge graph and the genetic algorithm-optimized fraud identification model, it is possible to accurately capture and identify the hidden associations and behavior patterns of fraud groups, thereby improving the accuracy of identification.

[0064] Strengthening the identification of complex fraud gangs: This invention not only focuses on individual abnormal behaviors, but also further models and analyzes the complex relationships between gangs through knowledge graphs, significantly enhancing the recognition of complex fraud behaviors and effectively reducing misjudgments and omissions.

[0065] Realize real-time monitoring and early warning of financial transactions: The system can monitor transaction activities in real time and continuously track suspected fraudulent transaction accounts. Once these accounts conduct transactions again, the system will immediately issue an early warning and terminate the transaction, thereby effectively preventing fraudulent behavior.

[0066] Improve anti-fraud execution efficiency and accuracy: With the help of a preset anti-fraud execution action library, the system can respond quickly based on feedback, such as freezing accounts suspected of fraud and notifying relevant departments for in-depth investigation, which greatly improves processing efficiency and accuracy.

[0067] Enhance the scalability and maintainability of the system: This technical solution is easy to deploy and upgrade, and based on advanced knowledge graphs and machine learning algorithms, it ensures high scalability and easy maintenance of the system.

[0068] Optimize user experience and trust: Through timely warning and fraudulent transaction blocking, it not only reduces the economic and psychological losses suffered by users due to fraud, but also significantly improves users' trust in the financial transaction system and overall user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0070] Figure 1 A flow chart of a method for identifying fraud gangs based on knowledge graph provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0071] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. The technical solutions provided by the embodiments of the present invention are described in detail below in conjunction with the drawings.

[0072] See also Figure 1 The present invention provides a method for identifying fraud gangs based on knowledge graph, comprising:

[0073] The present invention provides a method for identifying fraud gangs based on knowledge graph, comprising:

[0074] Step S101, the server side continuously obtains the user's identity authentication data in financial transactions and the user's normal transaction data from the application and the website;

[0075] Step S102, constructing a fraud identification model based on historical fraud data and normal transaction data, and optimizing the fraud identification model using a genetic algorithm to obtain an optimized fraud identification model;

[0076] Step S103: continuously obtain the user's identity verification data in financial transactions and the user's normal transaction data from the application and the website and substitute them into the optimized fraud identification model to obtain identification result information;

[0077] Step S104, determining the transaction account information according to the identity verification data in the identification result information, and when the transaction account information corresponding to the identification result information is traded again, generating an early warning message and sending it to the backend control terminal and the user terminal, and stopping the transaction execution corresponding to the transaction account information;

[0078] Step S105, executing a preset anti-fraud execution action according to the information fed back by the backend control terminal and the user terminal.

[0079] Specifically, the method for identifying fraud gangs based on knowledge graph provided by the present invention, step S101, the server side continuously obtains the user's identity authentication data and the user's normal transaction data in financial transactions from the application and the website, including:

[0080] The server side first needs to set up clear data collection interfaces, which should be compatible with the data output formats of applications and websites to ensure smooth data transmission.

[0081] Establish specifications for data collection and clarify which data is necessary, such as user authentication information (such as username, password hash, dynamic token, etc.), transaction information (such as transaction amount, transaction time, counterparty account, etc.), and device information (such as IP address, device ID, operating system, etc.).

[0082] Set the frequency of data collection to ensure that all important data is captured without affecting the performance of applications and websites.

[0083] The server side continuously fetches data from applications and websites through the set interface. This process needs to ensure the integrity and accuracy of the data.

[0084] Perform preliminary verification on the captured data to check whether the data format is correct and whether there are missing values ​​or outliers to ensure data quality.

[0085] The verified data is stored in the server's database for subsequent analysis.

[0086] Preprocess the collected raw data, including data cleaning (removing duplicate, invalid or erroneous data) and data conversion (converting data into a format suitable for analysis).

[0087] Integrate identity verification data and transaction data to establish connections between the data and form a complete data set. For example, match each transaction with its corresponding identity verification information.

[0088] Further quality checks are carried out on the integrated data to ensure its accuracy and consistency, providing a high-quality data foundation for subsequent identification of fraud gangs.

[0089] Specifically, the fraud group identification method based on knowledge graph provided by the present invention builds a fraud identification model based on historical fraud data and normal transaction data, and optimizes the fraud identification model using a genetic algorithm to obtain an optimized fraud identification model, including:

[0090] The server first organizes and prepares historical fraud data and normal transaction data to ensure the quality and integrity of the data.

[0091] Perform feature engineering to extract meaningful features from historical data, including transaction frequency, transaction amount distribution, transaction time pattern, user behavior pattern, etc.

[0092] The extracted features are selected and transformed to remove redundant and irrelevant information. At the same time, some techniques such as normalization and standardization are used to process the features to make them more suitable for model training.

[0093] Select a suitable machine learning algorithm or model framework to build an initial fraud identification model, such as decision tree, random forest, neural network and other algorithms.

[0094] Use the prepared dataset to train the model and optimize the performance of the model by adjusting model parameters and learning strategies.

[0095] Preliminarily evaluate the performance of the model, use techniques such as cross-validation to ensure that the model is not overfitting or underfitting, and record the performance metrics of the initial model.

[0096] Define the fitness function of the genetic algorithm, which should be able to reflect the performance of the fraud identification model. For example, accuracy, recall rate, F1 score, etc. can be used as evaluation indicators.

[0097] Initialize the population of the genetic algorithm, where each individual represents a set of parameter configurations for the fraud identification model.

[0098] The optimization process of the genetic algorithm, including operations such as selection, crossover, and mutation, is carried out to find the optimal model parameter configuration through continuous iteration.

[0099] After each generation iteration, the fitness of individuals in the population is evaluated, individuals with high fitness are selected to enter the next generation, and individuals with low fitness are eliminated.

[0100] When the preset number of iterations is reached or other stopping conditions are met, the optimization process of the genetic algorithm is stopped and the optimized fraud recognition model parameter configuration is output.

[0101] Specifically, the fraud group identification method based on knowledge graph provided by the present invention continuously obtains the user's identity authentication data in financial transactions and the user's normal transaction data from applications and websites and substitutes them into the optimized fraud identification model to obtain identification result information, including:

[0102] The server side continuously obtains the latest user identity authentication data and normal transaction data in financial transactions from applications and websites through the previously set interface.

[0103] Preprocess the newly acquired data to ensure that the format and structure of the data are consistent with the data used in model training, including operations such as data cleaning, conversion, and standardization.

[0104] Verify the completeness and validity of the data and remove any incomplete or invalid data records to ensure the quality of data input into the model.

[0105] Substitute the preprocessed data into the fraud identification model previously optimized by genetic algorithm.

[0106] Run the model to process and analyze the data. The model will determine whether there is a risk of fraud for each transaction based on its internal algorithms and logic.

[0107] Get the output of the model, which typically includes a fraud risk score for each transaction, a predicted label (such as fraudulent or not fraudulent), and other relevant information.

[0108] Parse the output of the model and convert it into a format that can be understood by the server and subsequent processing systems.

[0109] The parsed recognition result information is stored in the server's database for subsequent query and analysis.

[0110] As needed, the identification result information is sent to relevant systems or personnel for further processing or verification.

[0111] Specifically, the fraud gang identification method based on knowledge graph provided by the present invention determines the transaction account information according to the identity authentication data in the identification result information, and when the transaction account information corresponding to the identification result information is traded again, generates warning information and sends it to the background control terminal and the user terminal, and stops the transaction execution corresponding to the transaction account information, including:

[0112] The server performs an in-depth analysis of the recognition result information obtained in step S103, paying particular attention to transactions that are marked by the model as high-risk or suspected fraud.

[0113] Extract identity verification data from high-risk transactions and use this data to accurately determine the account information involved in the transaction.

[0114] Establish a monitoring list and add those trading accounts identified as suspected fraud to the list for subsequent monitoring.

[0115] A real-time monitoring mechanism is set up on the server side to continuously monitor all financial transactions, especially the transaction activities of accounts in the monitoring list.

[0116] When an account in the monitoring list attempts to trade again, the server immediately triggers an early warning mechanism.

[0117] Prepare to generate early warning information, which should include trading account number, transaction details, risk warnings, and recommended actions.

[0118] The server sends the warning information to the backend control and user ends in real time to ensure that relevant personnel and users can know and take action immediately.

[0119] At the same time, the server automatically executes transaction blocking operations to suspend or cancel current transactions of accounts in the monitoring list to prevent potential fraudulent activities from being completed.

[0120] Record the detailed information of the warning and transaction blocking, including time, account number, transaction details, etc., for subsequent audit and analysis.

[0121] Specifically, the fraud gang identification method based on knowledge graph provided by the present invention, step S105, executes a preset anti-fraud execution action according to the information fed back by the background control terminal and the user terminal, including:

[0122] The server side sets up a receiving mechanism for obtaining feedback information from the background control side and the user side in real time.

[0123] When feedback information arrives, the server immediately receives and parses it to ensure that the content and intent of the feedback are accurately understood.

[0124] Based on the results of the analysis, the feedback information is classified into categories such as confirmed fraud, false alarm, need for further investigation, etc.

[0125] According to the classification of the feedback information, the server queries the preset anti-fraud execution action library, which contains a series of pre-defined execution actions for different situations.

[0126] Determine the execution actions that match the current feedback information, which may include freezing the account suspected of fraud, notifying relevant departments for further investigation, lifting restrictions on false positive accounts, etc.

[0127] If the feedback information indicates that further investigation or human intervention is required, the server will trigger the corresponding process, such as transferring the case to a dedicated investigation team.

[0128] The server automatically executes certain anti-fraud actions, such as freezing accounts, sending notifications, etc.

[0129] While executing the action, the server will generate detailed execution records, including execution time, execution action type, account information involved, etc. These records will be stored in the server's log system for subsequent auditing, analysis and improvement of the anti-fraud system.

[0130] The present invention effectively solves the problem of lack of comprehensive analysis and identification means in identifying financial fraud gangs through comprehensive technical means.

[0131] First, through continuous data acquisition and integration, the system can continuously collect users' identity authentication data and normal transaction data in financial transactions from applications and websites. These data are initially verified, stored, pre-processed and integrated to ensure the quality and integrity of the data, providing a rich data foundation for subsequent fraud gang identification.

[0132] Secondly, the present invention uses historical fraud data and normal transaction data to build a fraud identification model, and optimizes it through a genetic algorithm to improve the accuracy and efficiency of the model. This enables the system to identify fraudulent behavior more accurately.

[0133] Furthermore, the real-time transaction monitoring and early warning mechanism enables the system to promptly detect and respond to suspected fraudulent transactions. When the transaction account in the identification result information conducts transactions again, the system will generate an early warning message and send it to the backend control terminal and the user terminal, and stop the execution of related transactions, thereby effectively preventing the occurrence of fraud.

[0134] In addition, the present invention further enhances the flexibility and adaptability of the system by executing preset anti-fraud actions according to feedback information from the background control terminal and the user terminal through feedback information processing and anti-fraud execution mechanism.

[0135] In summary, the present invention effectively solves the problem of lack of comprehensive analysis and identification means in identifying financial fraud gangs through various technical means.

[0136] Those skilled in the art can clearly understand that the technology in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution in the embodiments of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention or some parts of the embodiments.

[0137] The above-described embodiments of the present invention do not limit the protection scope of the present invention.

Claims

1. A method for identifying fraud gangs based on knowledge graph, characterized in that: include; Step S101, the server side continuously obtains the user's identity authentication data in financial transactions and the user's normal transaction data from the application and the website; Step S102, constructing a fraud identification model based on historical fraud data and normal transaction data, and optimizing the fraud identification model using a genetic algorithm to obtain an optimized fraud identification model; Step S103: continuously obtain the user's identity verification data in financial transactions and the user's normal transaction data from the application and the website and substitute them into the optimized fraud identification model to obtain identification result information; Step S104, determining the transaction account information according to the identity authentication data in the identification result information, and when the transaction account information corresponding to the identification result information is traded again, generating an early warning message and sending it to the backend control terminal and the user terminal, and stopping the transaction execution corresponding to the transaction account information; Step S105, executing a preset anti-fraud execution action according to the information fed back by the backend control terminal and the user terminal.

2. The method for identifying fraud gangs based on knowledge graph according to claim 1, characterized in that: Step S101, the server side continuously obtains the user's identity authentication data in financial transactions and the user's normal transaction data from the application and the website, including: The server sets a clear data collection interface and frequency. The server continuously obtains data from applications and websites through the set interface. Perform preliminary verification on the captured data to check whether the data format is correct, whether there are missing values ​​or abnormal values, and store the verified data in the server database; Pre-process the collected raw data, integrate the identity verification data and transaction data, establish the connection between the data, and form a complete data set.

3. The method according to claim 1, characterized in that Step S102, constructing a fraud identification model based on historical fraud data and normal transaction data, and optimizing the fraud identification model using a genetic algorithm to obtain an optimized fraud identification model, including: The server organizes and prepares historical fraud data and normal transaction data, and extracts features from the historical data, including transaction frequency, transaction amount distribution, transaction time pattern, and user behavior pattern; Select and transform the extracted features, choose the appropriate machine learning algorithm to build the initial fraud detection model, use the prepared dataset to train the model, and optimize the model's performance by adjusting the model parameters and learning strategies; Define the fitness function of the genetic algorithm, which should be able to reflect the performance of the fraud identification model, and perform the optimization process of the genetic algorithm, including operations such as selection, crossover, and mutation, to find the optimal model parameter configuration through iteration; When the preset number of iterations is reached or other stopping conditions are met, the optimization process of the genetic algorithm is stopped and the optimized fraud recognition model parameter configuration is output.

4. The method according to claim 1, characterized in that Step S103, continuously obtaining the user's identity verification data in financial transactions and the user's normal transaction data from the application and the website and substituting them into the optimized fraud identification model to obtain identification result information, including: The server side continuously obtains the latest user identity authentication data and normal transaction data in financial transactions from applications and websites through the set interface; Preprocess the newly acquired data to ensure that the format and structure of the data are consistent with the data used in model training, verify the integrity and validity of the data, and eliminate any incomplete or invalid data records; Substitute the preprocessed data into the fraud identification model previously optimized by genetic algorithm; Run the model to process and analyze the data. The model will determine whether there is a fraud risk for each transaction based on its internal algorithms and logic. Get the model output, which usually includes the fraud risk score and predicted label for each transaction. The output results of the model are parsed and converted into a format that matches the server and subsequent processing system, and the parsed recognition result information is stored in the server's database.

5. The method according to claim 1, characterized in that Step S104, determining the transaction account information according to the identity authentication data in the identification result information, and when the transaction account information corresponding to the identification result information is traded again, generating an early warning message and sending it to the backend control terminal and the user terminal, and stopping the transaction execution corresponding to the transaction account information, including; The server performs an in-depth analysis of the high-risk or suspected fraudulent transactions in the identification result information obtained in step S103, extracts identity verification data from the high-risk transactions, and determines the transaction account information involved; Establish a monitoring list and add the trading accounts identified as suspected fraud to the list. Set up a real-time monitoring mechanism on the server side to continuously monitor the trading activities of the accounts in the monitoring list. When an account in the monitoring list attempts to trade again, the server immediately triggers the early warning mechanism and generates early warning information, which includes the trading account, transaction details, risk warnings, and recommended actions.

6. The method according to claim 1, characterized in that Step S105, executing a preset anti-fraud execution action according to the information fed back by the backend control terminal and the user terminal, including: The server side sets up a receiving mechanism to obtain feedback information from the backend control side and the user side in real time; When feedback information arrives, the server immediately receives and parses the information, and classifies the feedback information according to the parsing results. The classification of feedback information includes confirmed fraud, false alarm, and need for further investigation; According to the classification of the feedback information, the server queries a preset anti-fraud execution action library, which includes pre-defined execution actions for different situations; Determine the execution actions that match the current feedback information, including freezing the suspected fraud account, notifying relevant departments for further investigation, and lifting the false positive account.

Images