A payment transaction information processing method and system

Abstract

The present invention discloses a payment transaction information processing method and system, which relates to the technical field of transaction information processing, including setting an initial time window based on account data and performing identity authentication, continuously collecting user facial features, and performing periodic facial recognition and verification; setting a sub-time window based on the payment transaction frequency, performing data aggregation and forming a comprehensive verification set, performing abnormality scoring through transaction information, and analyzing and determining transaction risks. The method of the present invention can set a more personalized initial time window by analyzing the average transaction frequency of different user account types, and improve the utilization efficiency of facial information through periodic facial information based on the initial time window. By setting a sub-time window, the system can perform fine-grained monitoring in a shorter time period, and aggregate the transaction data in the sub-time window and then perform unified verification for abnormality scoring, reducing repeated verification requests for each transaction.

Description

Technical Field

[0001] The present invention relates to the technical field of transaction information processing, and in particular to a payment transaction information processing method and system. Background Art

[0002] With the rapid development of e-commerce and the popularization of Internet finance, payment transaction information processing technology has made significant progress in recent years. Traditional payment transaction systems mainly rely on fixed verification methods, such as passwords, one-time verification codes, etc. These methods have effectively ensured the security of transactions in the past period of time. However, with the increase in transaction frequency and transaction amount, as well as the increasing complexity of online fraud and network attacks, a single verification method can no longer meet the security requirements of modern payment systems. For this reason, the emergence of some biometric verification systems has further increased the security of payment transaction information.

[0003] Although existing payment transaction information processing technology has achieved certain development, when faced with high-frequency transactions, risk assessment is often based on static rules and preset thresholds, which makes it inconvenient to make full use of transaction information for dynamic adjustment. It is easy for the verification frequency to be too high, resulting in reduced transaction efficiency and poor user experience. It is inconvenient to dynamically adjust the transaction verification frequency based on transaction risks, resulting in insufficient security verification for high-risk transactions and too frequent security verification for low-risk transactions, which reduces the transaction efficiency of users. Summary of the invention

[0004] In view of the problems existing in the above-mentioned existing payment transaction information processing method and system, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is that when faced with high-frequency transactions, risk assessment is often performed based on static rules and preset thresholds, which makes it inconvenient to make full use of transaction information for dynamic adjustment, and easily leads to a situation where the verification frequency is too high, resulting in reduced transaction efficiency and poor user experience. It is inconvenient to dynamically adjust the transaction verification frequency based on transaction risks, resulting in insufficient security verification for high-risk transactions and too frequent security verification for low-risk transactions, which reduces the transaction efficiency of users.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a payment transaction information processing method, comprising:

[0007] Set an initial time window based on account data and perform identity authentication, continuously collect user facial features, and perform periodic facial recognition and verification;

[0008] Set sub-time windows based on the payment transaction frequency, aggregate data and form a comprehensive verification set, perform anomaly scoring based on transaction information, and analyze and determine transaction risks;

[0009] The anomaly score of the initial time window is integrated to perform the final risk assessment and dynamically adjust the initial time window;

[0010] Transaction data is securely transmitted and stored based on information encryption technology.

[0011] As a preferred solution of the payment transaction information processing method of the present invention, wherein: the initial time window is set based on the account data and identity authentication is performed, including:

[0012] Based on the account database and transaction record database, extract user account type data and user historical transaction data, the user historical transaction data includes transaction frequency, transaction amount, and transaction time;

[0013] Conduct statistical analysis on historical transaction data and calculate the average transaction frequency of different user account types;

[0014] The initial time window T for transactions is based on the user account type and average transaction frequency.

[0015] As a preferred solution of the payment transaction information processing method of the present invention, the continuous collection of user facial features and periodic facial recognition and verification include:

[0016] Biometric verification is performed when users initiate transaction requests;

[0017] Biometric verification is based on the face recognition system. The facial feature data in the account database is compared with the user's current facial feature data collected by the camera to perform facial recognition. Through dynamic password verification, a one-time dynamic password is sent to the user's registered mobile phone number for password recognition to complete preliminary identity verification;

[0018] Based on the initial time window T, the user's facial feature data is continuously collected through the camera and the data is cached;

[0019] According to the average transaction frequency of historical transaction data, the periodic verification time is set, expressed as:

[0020] ;

[0021] in represents the periodic verification time, and f represents the average transaction frequency;

[0022] Within the initial time window T The face recognition system performs periodic face recognition based on the cached facial feature data at a certain time.

[0023] If facial recognition passes, the cached facial feature data is updated and the next periodic verification time is continued;

[0024] If the verification fails, the user is prompted to re-authenticate.

[0025] As a preferred solution of the payment transaction information processing method of the present invention, wherein: the sub-time window is set based on the payment transaction frequency, data aggregation is performed and a comprehensive verification set is formed, including:

[0026] Within the set initial time window T, real-time monitoring is performed based on the user's payment transaction request, including transaction time, transaction amount and transaction location;

[0027] Based on the payment transaction frequency within the initial time window T, a sub-time window is set, expressed as:

[0028] ;

[0029] ;

[0030] in represents the i-th sub-time window, n represents the number of sub-time windows, represents the expected average number of transactions in each sub-time window, Indicates the historical transaction frequency;

[0031] The data of transaction requests within the sub-time window are aggregated to form a comprehensive verification set, including the transaction time, transaction amount and transaction location of the transaction request and the preliminary identity verification result, and the aggregated comprehensive verification request is sent to the verification server.

[0032] As a preferred solution of the payment transaction information processing method of the present invention, wherein: the abnormal scoring through transaction information, analyzing and determining transaction risks includes:

[0033] The verification server verifies the transaction data in the comprehensive verification request, where the amount of each transaction request is verified, expressed as:

[0034] ;

[0035] in Indicates the current transaction amount. Indicates the historical average transaction amount of the user account. Indicates the standard deviation of the historical transaction amount of the user account, Indicates the deviation amount;

[0036] The location of each transaction request is verified as:

[0037] ;

[0038] in and Indicates the coordinates of the current transaction location. and represents the coordinates of historical commonly used transaction locations, and D represents the verification distance;

[0039] The time of each transaction request is verified, expressed as:

[0040] ;

[0041] in Indicates the deviation time, Indicates the current trading time. Indicates regular trading hours;

[0042] Amount of deviation , verify distance D and deviation time The standard deviation is processed separately, and a linear regression model is constructed. The standardized transaction amount, transaction location and transaction time are comprehensively considered and anomaly scores are performed, which are expressed as:

[0043] ;

[0044] in represents the anomaly score of the ith comprehensive validation set, , and are weight coefficients, represents the standardized deviation amount, represents the normalized verification distance, represents the normalized deviation time;

[0045] Use the training set to calculate the loss through the mean square error loss function, use the Adam optimizer and the gradient descent method to iteratively optimize the model parameters. When the loss of the linear regression model no longer decreases significantly during the continuous iteration process, stop iterating and output the model parameters to update the linear regression model.

[0046] Based on the 99.7% coverage principle, the sum of the mean of historical anomaly scores and three times the standard deviation is used as the transaction threshold. If the anomaly score S calculated by the aggregated comprehensive verification data is greater than or equal to the transaction threshold, it is judged that there is a transaction risk. If the anomaly score S calculated by the aggregated comprehensive verification data is less than the transaction threshold, it is judged that there is no transaction risk.

[0047] As a preferred solution of the payment transaction information processing method of the present invention, wherein: the abnormal score of the integrated initial time window is used to perform a final risk assessment, and the initial time window is dynamically adjusted, including:

[0048] Based on the abnormal score value of each sub-time window within the initial time window T as the scoring feature X, a machine learning model is constructed for the final risk assessment, which is expressed as:

[0049] ;

[0050] Where R represents the final risk score, represents the sigmoid function, w represents the model weight, and b represents the bias term;

[0051] The loss is calculated using the loss function on the training set, and the model parameters are iteratively optimized using the Adam optimizer and the gradient descent method. When the loss of the machine learning model no longer decreases significantly during the continuous iteration process, the iteration is stopped and the model parameters are output to update the machine learning model.

[0052] Calculating the Historical Final Risk Score The mean and standard deviation of the historical final risk score The sum of the mean and standard deviation of is used as the high risk threshold to set the historical final risk score The difference between the mean and standard deviation is used as the low risk threshold;

[0053] If the final risk score R is greater than or equal to the low risk threshold and less than the high risk threshold, the risk is determined to be moderate; if the final risk score R is greater than or equal to the high risk threshold, the transaction is determined to be high risk and the initial time window is shortened; if the final risk score R is less than the low risk threshold, the transaction is determined to be low risk and the initial time window is extended;

[0054] Adjusting the initial time window is expressed as:

[0055] ;

[0056] ;

[0057] in represents the initial time window for adjustment, represents the adjustment coefficient, sig represents the sigmiod function, represents the maximum value of the historical final risk score, represents the minimum value of the historical final risk score, represents the standard deviation of the historical terminal risk score, represents the mean of the historical terminal risk score, and sign represents the sign function.

[0058] As a preferred solution of the payment transaction information processing method described in the present invention, wherein: the secure transmission and storage of transaction data based on information encryption technology refers to the use of the advanced encryption standard AES to encrypt the cached verification data and aggregated transaction data during the data transmission process, and the use of the secure transmission protocol TLS for secure data transmission. When storing the data in the account database and the transaction record database, end-to-end encryption is performed and transparent data encryption TDE is used for encryption.

[0059] Another object of the present invention is to provide a system for a payment transaction information processing method, which comprises:

[0060] The authentication module maintains user account information, sets the initial time window, and performs identity authentication;

[0061] The periodic verification submodule continuously collects user facial feature data and performs periodic facial recognition and verification;

[0062] The data processing and verification module sets sub-time windows based on the payment transaction frequency, aggregates data and performs unified verification, scores abnormalities based on transaction information, and analyzes and determines transaction risks.

[0063] The risk assessment module combines the anomaly scores of the initial time window to perform the final risk assessment;

[0064] The time window adjustment module dynamically adjusts the initial time window according to the final risk score;

[0065] The secure transmission module performs secure transmission and storage of data based on information processing.

[0066] A computer device comprises: a memory and a processor; the memory stores a computer program, and the processor implements the steps of the above-mentioned payment transaction information processing method when executing the computer program.

[0067] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the payment transaction information processing method.

[0068] The beneficial effects of the present invention are as follows: by analyzing the average transaction frequency of different user account types, a more personalized initial time window can be set, and based on the initial time window, periodic facial information is used to improve the utilization efficiency of facial information. By setting a sub-time window, the system can perform fine-grained monitoring in a shorter time period, and aggregate the transaction data in the sub-time window before unified verification for abnormal scoring, thereby reducing repeated verification requests for each transaction and optimizing the performance of the system. By aggregating abnormal score values ​​into features for unified evaluation, the overall processing efficiency can be improved, and when the initial time window is dynamically adjusted, system resources can be effectively allocated, so that high-risk transactions receive more attention and processing, thereby improving the smoothness of transactions and user satisfaction, and being promptly identified and processed when high-risk transactions are determined, thereby enhancing security. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0070] Figure 1 A flowchart of a payment transaction information processing method.

[0071] Figure 2 A schematic diagram of the structure of a payment transaction information processing system. DETAILED DESCRIPTION

[0072] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0073] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0074] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0075] Example 1, reference Figure 1, which is the first embodiment of the present invention, and provides a payment transaction information processing method, the payment transaction information processing method comprising:

[0076] S1, sets the initial time window based on account data and performs identity authentication, continuously collects user facial features, and performs periodic facial recognition and verification;

[0077] Preferably, setting an initial time window based on account data and performing identity authentication includes,

[0078] Based on the account database and transaction record database, extract user account type data and user historical transaction data, the user historical transaction data includes transaction frequency, transaction amount, and transaction time;

[0079] Conduct statistical analysis on historical transaction data and calculate the average transaction frequency of different user account types;

[0080] The initial time window T for transactions is based on the user account type and average transaction frequency.

[0081] By analyzing the average transaction frequency of different user account types, a more personalized initial time window can be set to improve user experience and system efficiency. For users with higher transaction frequencies, the system can set a shorter initial time window to conduct risk assessment and monitoring more frequently, and promptly identify and deal with potential risks. For users with lower transaction frequencies, the system can set a longer initial time window to reduce unnecessary waste of resources and improve system processing efficiency. Based on the user's average transaction frequency, the system can more effectively detect abnormal transaction behavior. The personalized initial time window setting can significantly improve the user experience, allowing the system to respond to users' trading needs more flexibly and agilely.

[0082] Furthermore, the user's facial features are continuously collected for periodic facial recognition and verification, including:

[0083] Biometric verification is performed when users initiate transaction requests;

[0084] Biometric verification is based on the face recognition system. The facial feature data in the account database is compared with the user's current facial feature data collected by the camera to perform facial recognition. Through dynamic password verification, a one-time dynamic password is sent to the user's registered mobile phone number for password recognition to complete preliminary identity verification;

[0085] Based on the initial time window T, the user's facial feature data is continuously collected through the camera and the data is cached;

[0086] According to the average transaction frequency of historical transaction data, the periodic verification time is set, expressed as:

[0087] ;

[0088] in represents the periodic verification time, and f represents the average transaction frequency;

[0089] Within the initial time window T The face recognition system performs periodic face recognition based on the cached facial feature data at a certain time.

[0090] If facial recognition passes, the cached facial feature data is updated and the next periodic verification time is continued;

[0091] If the verification fails, the user is prompted to re-authenticate.

[0092] By combining face recognition and one-time dynamic password verification, it provides multi-level security protection, greatly increasing the security of the transaction process. The camera collects and caches facial feature data in real time to ensure continuous monitoring and verification during the transaction process to prevent identity fraud. The verification time is dynamically adjusted according to the user's transaction habits, which ensures security without affecting the user's normal operation. By caching facial feature data, the computational burden of repeated collection and comparison is reduced, the system efficiency is improved, and the situation where others take over the operation after the initial verification is prevented. The cache verification is performed based on the transaction person's facial information, which improves the utilization efficiency of facial information and improves the security and processing efficiency of payment transactions.

[0093] S2, set sub-time windows based on the payment transaction frequency, aggregate data and form a comprehensive verification set, perform anomaly scoring based on transaction information, and analyze and determine transaction risks;

[0094] Preferably, a sub-time window is set based on the payment transaction frequency, data is aggregated and a comprehensive verification set is formed, including:

[0095] Within the set initial time window T, real-time monitoring is performed based on the user's payment transaction request, including transaction time, transaction amount and transaction location;

[0096] Based on the payment transaction frequency within the initial time window T, a sub-time window is set, expressed as:

[0097] ;

[0098] ;

[0099] in represents the i-th sub-time window, n represents the number of sub-time windows, Represents the expected average number of transactions in each sub-time window. The expectation in the sub-time window is determined by calculating the sum of the mean and standard deviation of the historical transaction counts. Indicates the historical transaction frequency;

[0100] The data of transaction requests within the sub-time window are aggregated to form a comprehensive verification set, including the transaction time, transaction amount and transaction location of the transaction request and the preliminary identity verification result, and the aggregated comprehensive verification request is sent to the verification server.

[0101] By setting sub-time windows, the system can perform fine-grained monitoring in a shorter period of time, improve the ability to identify and respond to abnormal transactions, and conduct comprehensive verification based on transaction data within the sub-time windows. It can promptly discover and deal with potential risks and ensure the security of transactions. Transaction data within the sub-time windows are aggregated and then verified uniformly, which reduces repeated verification requests for each transaction and optimizes system performance. Through the comprehensive verification set, the verification server can process transaction requests more efficiently, reduce system delays, and improve overall processing efficiency.

[0102] Furthermore, the transaction information is used to perform abnormal scoring, analyze and determine transaction risks, including:

[0103] The verification server verifies the transaction data in the comprehensive verification request, where the amount of each transaction request is verified, expressed as:

[0104] ;

[0105] in Indicates the current transaction amount. Indicates the historical average transaction amount of the user account. Indicates the standard deviation of the historical transaction amount of the user account, Indicates the deviation amount;

[0106] The location of each transaction request is verified as:

[0107] ;

[0108] in and Indicates the coordinates of the current transaction location. and represents the coordinates of historical commonly used transaction locations, and D represents the verification distance;

[0109] The time of each transaction request is verified, expressed as:

[0110] ;

[0111] in Indicates the deviation time, Indicates the current trading time. Indicates the regular trading time period, which is defined as the regular trading time period based on the average start time and end time of the user's historical trading time;

[0112] Amount of deviation , verify distance D and deviation time The standard deviation is processed separately, and a linear regression model is constructed. The standardized transaction amount, transaction location and transaction time are comprehensively considered and anomaly scores are performed, which are expressed as:

[0113] ;

[0114] in represents the anomaly score of the ith comprehensive validation set, , and are weight coefficients, represents the standardized deviation amount, represents the normalized verification distance, represents the normalized deviation time;

[0115] The loss is calculated using the mean square error loss function using the training set, expressed as:

[0116] ;

[0117] Where L represents the calculation loss, J represents the number of samples, Indicates the actual value, Represents the model output;

[0118] The model parameters are iteratively optimized using the Adam optimizer and the gradient descent method. When the loss of the linear regression model no longer decreases significantly during the continuous iteration, the iteration is stopped and the model parameters are output to update the linear regression model.

[0119] Based on the 99.7% coverage principle, the sum of the mean of historical anomaly scores and three times the standard deviation is used as the transaction threshold. If the anomaly score S calculated by the aggregated comprehensive verification data is greater than or equal to the transaction threshold, it is judged that there is a transaction risk. If the anomaly score S calculated by the aggregated comprehensive verification data is less than the transaction threshold, it is judged that there is no transaction risk.

[0120] By verifying the transaction amount, transaction location and transaction time, the system can comprehensively evaluate the legality of each transaction and improve transaction security. Through standardized processing and linear regression models, it can accurately identify abnormal transactions and reduce the risks of false positives and negative reports. The transaction data within the sub-time window is aggregated and then uniformly verified, which reduces repeated calculations and optimizes system performance. The processing of the comprehensive verification set can more efficiently verify transactions and improve the system response speed. The system can dynamically adjust the verification strategy based on historical data to adapt to the transaction behaviors and risk levels of different users. Through accurate risk assessment, the system can reduce interference to users and improve the transaction experience.

[0121] S3, conducts final risk assessment based on the anomaly score of the initial time window and dynamically adjusts the initial time window;

[0122] Preferably, the final risk assessment is performed by integrating the abnormal scores of the initial time window and dynamically adjusting the initial time window, including:

[0123] Based on the abnormal score value of each sub-time window within the initial time window T as the scoring feature X, a machine learning model is constructed for the final risk assessment, which is expressed as:

[0124] ;

[0125] Where R represents the final risk score, represents the sigmoid function, w represents the model weight, and b represents the bias term;

[0126] The loss is calculated using the loss function through the training set, expressed as:

[0127] M =− 1 V ∑ i = 1 V [ r i log ( r ^ i ) + ( 1 − r i ) log ( 1 − r ^ i ) ] ;

[0128] Where M represents the calculation loss, V represents the total number of samples, represents the true label, Represents the model output value;

[0129] The model parameters are iteratively optimized using the Adam optimizer and the gradient descent method. When the loss of the machine learning model no longer decreases significantly during the continuous iteration, the iteration is stopped and the model parameters are output to update the machine learning model.

[0130] Calculating the Historical Final Risk Score The mean and standard deviation of the historical final risk score The sum of the mean and standard deviation of is used as the high risk threshold to set the historical final risk score The difference between the mean and the standard deviation was used as the low risk threshold;

[0131] If the final risk score R is greater than or equal to the low risk threshold and less than the high risk threshold, the risk is determined to be moderate; if the final risk score R is greater than or equal to the high risk threshold, the transaction is determined to be high risk and the initial time window is shortened; if the final risk score R is less than the low risk threshold, the transaction is determined to be low risk and the initial time window is extended;

[0132] Adjusting the initial time window is expressed as:

[0133] ;

[0134] ;

[0135] in represents the initial time window for adjustment, represents the adjustment coefficient, sig represents the sigmiod function, represents the maximum value of the historical final risk score, represents the minimum value of the historical final risk score, represents the standard deviation of the historical terminal risk score, It represents the mean of the historical terminal risk score, and sign represents the sign function, which is used to return positive and negative values.

[0136] By constructing a machine learning model based on the abnormal score value of each sub-time window within the initial time window as a feature, a comprehensive risk assessment of the transaction can be performed. The initial time window can be dynamically adjusted according to the final risk score, which can quickly adapt to changes in the trading environment, respond to potential risks in a timely manner, and improve system security. By aggregating the abnormal score values ​​as features for unified evaluation, the overall processing efficiency can be improved, and when the initial time window is dynamically adjusted, system resources can be effectively allocated, so that high-risk transactions receive more attention and processing, and low-risk transactions reduce unnecessary interference. The system dynamically adjusts the initial time window according to the risk score, which can reduce frequent verification of low-risk users, improve the smoothness of transactions and user satisfaction, and timely identify and process high-risk transactions when they are determined, enhancing users' trust in the security of the platform. By analyzing and learning historical data, the system can continuously optimize the risk assessment model and the initial time window adjustment strategy to improve overall risk management capabilities.

[0137] S4, securely transmit and store transaction data based on information encryption technology;

[0138] Preferably, secure transmission and storage of transaction data based on information encryption technology means that during the data transmission process, the cached verification data and aggregated transaction data are encrypted using the advanced encryption standard AES, and secure data transmission is performed using the secure transmission protocol TLS. When the data is stored in the account database and transaction record database, end-to-end encryption is performed and transparent data encryption TDE is used for encryption.

[0139] By using the Advanced Encryption Standard (AES) to encrypt cached verification data and transaction data, the security of data during transmission and storage is ensured. End-to-end encryption ensures that data remains encrypted from the sender to the receiver, preventing data leakage at intermediate nodes or paths. This achieves comprehensive protection of data in all aspects of transmission, processing and storage, reducing the risk of data leakage and tampering.

[0140] Example 2, reference Figure 2 , which is the second embodiment of the present invention, and which is different from the previous embodiment, provides a system for a payment transaction information processing method, including:

[0141] The authentication module maintains user account information, sets the initial time window, and performs identity authentication;

[0142] The periodic verification submodule continuously collects user facial feature data and performs periodic facial recognition and verification;

[0143] The data processing and verification module sets sub-time windows based on the payment transaction frequency, aggregates data and performs unified verification, scores abnormalities based on transaction information, and analyzes and determines transaction risks.

[0144] The risk assessment module combines the anomaly scores of the initial time window to perform the final risk assessment;

[0145] The time window adjustment module dynamically adjusts the initial time window according to the final risk score;

[0146] The secure transmission module performs secure transmission and storage of data based on information processing.

[0147] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0148] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0149] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0150] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit with a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit with a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0151] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A payment transaction information processing method, characterized in that: include, Set an initial time window based on account data and perform identity authentication, continuously collect user facial features, and perform periodic facial recognition and verification; Set sub-time windows based on the payment transaction frequency, aggregate data and form a comprehensive verification set, perform anomaly scoring based on transaction information, and analyze and determine transaction risks; The anomaly score of the initial time window is integrated to perform the final risk assessment and dynamically adjust the initial time window; Securely transmit and store transaction data based on information encryption technology; The initial time window is set based on the account data and identity verification is performed, including: Based on the account database and transaction record database, extract user account type data and user historical transaction data, the user historical transaction data includes transaction frequency, transaction amount, and transaction time; Conduct statistical analysis on historical transaction data and calculate the average transaction frequency of different user account types; Determine the average authentication window time based on the user account type and average transaction frequency, and use it as the initial time window T for the transaction; The continuous collection of user facial features and periodic facial recognition and verification include: Biometric verification is performed when users initiate transaction requests; Biometric verification is based on the face recognition system. The facial feature data in the account database is compared with the user's current facial feature data collected by the camera to perform facial recognition. Through dynamic password verification, a one-time dynamic password is sent to the user's registered mobile phone number for password recognition to complete preliminary identity verification; Based on the initial time window T, the user's facial feature data is continuously collected through the camera and the data is cached; According to the average transaction frequency of historical transaction data, the periodic verification time is set, expressed as: ； in represents the periodic verification time, and f represents the average transaction frequency; Within the initial time window T The face recognition system performs periodic face recognition based on the cached facial feature data at a certain time. If facial recognition passes, the cached facial feature data is updated and the next periodic verification time is continued; If the verification fails, the user is prompted to re-authenticate. The sub-time window is set based on the payment transaction frequency, data is aggregated and a comprehensive verification set is formed, including: Within the set initial time window T, real-time monitoring is performed based on the user's payment transaction request, including transaction time, transaction amount and transaction location; Based on the payment transaction frequency within the initial time window T, a sub-time window is set, expressed as: ； ； in represents the i-th sub-time window, n represents the number of sub-time windows, represents the expected average number of transactions in each sub-time window, Indicates the historical transaction frequency; Aggregate the data of transaction requests within the sub-time window to form a comprehensive verification set, including the transaction time, transaction amount and transaction location of the transaction request and the preliminary identity verification result, and send the aggregated comprehensive verification request to the verification server; The abnormal scoring of transaction information, analysis and determination of transaction risks include: The verification server verifies the transaction data in the comprehensive verification request, where the amount of each transaction request is verified, expressed as: ； in Indicates the current transaction amount. Indicates the historical average transaction amount of the user account. Indicates the standard deviation of the historical transaction amount of the user account, Indicates the deviation amount; The location of each transaction request is verified as: ； in and Indicates the coordinates of the current transaction location. and represents the coordinates of historical commonly used transaction locations, and D represents the verification distance; The time of each transaction request is verified, expressed as: ； in Indicates the deviation time, Indicates the current trading time. Indicates regular trading hours; Amount of deviation , verify distance D and deviation time The standard deviation is processed separately, and a linear regression model is constructed. The standardized transaction amount, transaction location and transaction time are comprehensively considered and anomaly scores are performed, which are expressed as: ； in represents the anomaly score of the ith comprehensive validation set, , and are weight coefficients, represents the standardized deviation amount, represents the normalized verification distance, represents the normalized deviation time; Use the training set to calculate the loss through the mean square error loss function, use the Adam optimizer and the gradient descent method to iteratively optimize the model parameters. When the loss of the linear regression model no longer decreases significantly during the continuous iteration process, stop iterating and output the model parameters to update the linear regression model. Based on the 99.7% coverage principle, the sum of the mean of the historical anomaly score and three times the standard deviation is used as the transaction threshold. If the anomaly score S calculated by the aggregated comprehensive verification data is greater than or equal to the transaction threshold, it is determined that there is a transaction risk. If the anomaly score S calculated by the aggregated comprehensive verification data is less than the transaction threshold, it is determined that there is no transaction risk. The abnormal score of the integrated initial time window is used to perform a final risk assessment, and the initial time window is dynamically adjusted, including: Based on the abnormal score value of each sub-time window within the initial time window T as the scoring feature X, a machine learning model is constructed for the final risk assessment, which is expressed as: ； Where R represents the final risk score, represents the sigmoid function, w represents the model weight, and b represents the bias term; The loss is calculated using the loss function on the training set, and the model parameters are iteratively optimized using the Adam optimizer and the gradient descent method. When the loss of the machine learning model no longer decreases significantly during the continuous iteration process, the iteration is stopped and the model parameters are output to update the machine learning model. Calculating the Historical Final Risk Score The mean and standard deviation of the historical final risk score The sum of the mean and standard deviation of is used as the high risk threshold to set the historical final risk score The difference between the mean and standard deviation is used as the low risk threshold; If the final risk score R is greater than or equal to the low risk threshold and less than the high risk threshold, the risk is determined to be moderate; if the final risk score R is greater than or equal to the high risk threshold, the transaction is determined to be high risk and the initial time window is shortened; if the final risk score R is less than the low risk threshold, the transaction is determined to be low risk and the initial time window is extended; Adjusting the initial time window is expressed as: ； ； in represents the initial time window for adjustment, represents the adjustment coefficient, sig represents the sigmiod function, represents the maximum value of the historical final risk score, represents the minimum value of the historical final risk score, represents the standard deviation of the historical terminal risk score, represents the mean of the historical terminal risk score, and sign represents the sign function.

2. The payment transaction information processing method according to claim 1, characterized in that: The secure transmission and storage of transaction data based on information encryption technology refers to the use of the Advanced Encryption Standard AES to encrypt cached verification data and aggregated transaction data during data transmission, the use of the secure transmission protocol TLS for secure data transmission, and end-to-end encryption when storing data in the account database and transaction record database, and encryption using transparent data encryption TDE.

3. A system based on the payment transaction information processing method according to any one of claims 1-2, characterized in that: include, The authentication module maintains user account information, sets the initial time window, and performs identity authentication; The periodic verification submodule continuously collects user facial feature data and performs periodic facial recognition and verification; The data processing and verification module sets sub-time windows based on the payment transaction frequency, aggregates data and performs unified verification, scores abnormalities based on transaction information, and analyzes and determines transaction risks. The risk assessment module combines the anomaly scores of the initial time window to perform the final risk assessment; The time window adjustment module dynamically adjusts the initial time window according to the final risk score; The secure transmission module performs secure transmission and storage of data based on information processing.

4. A computer device comprising: A memory and a processor; the memory stores a computer program, wherein the processor implements the steps of the payment transaction information processing method described in any one of claims 1 to 2 when executing the computer program.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the payment transaction information processing method described in any one of claims 1 to 2 are implemented.

Images