Transaction risk identification method and equipment based on AI intelligence

Abstract

The invention discloses a transaction risk identification method and device based on AI intelligence, and relates to the technical field of risk identification, and the method comprises the steps: obtaining user transaction information and behavior data in real time, carrying out the preprocessing of the data, extracting data features, and generating a feature matrix; and generating a transaction risk score according to the feature matrix, carrying out preliminary risk classification, calculating a positive and negative ideal solution distance of the classified transaction data, and calculating the risk proximity of the transaction according to the ideal solution distance. Through a multi-stage classification algorithm and positive and negative ideal solution distance calculation, the accuracy and the refinement level of risk identification are remarkably improved, the transaction risk proximity is calculated by adopting the positive and negative ideal solution distance, quantitative sorting of transaction risks is realized, transaction characteristics can be deeply mined, and the transaction risk identification accuracy and the transaction risk identification precision are improved. And the risk assessment model is dynamically optimized in the full life cycle of the transaction, so that the transaction risk identification efficiency and accuracy in a complex transaction scene are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of risk identification technology, and in particular to a transaction risk identification method and device based on AI intelligence. Background Art

[0002] With the rapid development of e-commerce, online payment and financial technology, the transaction volume has shown an exponential growth, and the real-time and complexity of transactions are also increasing. However, the diversity and complexity of transaction scenarios have made traditional transaction risk identification methods face unprecedented challenges. In the existing technology, rule-based and statistical analysis methods can identify some risky transactions to a certain extent, but these methods rely on the setting of fixed rules and are difficult to deal with new fraudulent behaviors or abnormal patterns. In recent years, the rapid development of artificial intelligence (AI) technology has brought new opportunities to the field of transaction risk identification. Through machine learning and deep learning models, multi-dimensional features can be extracted from complex transaction data to achieve more efficient risk assessment. In the existing technology, traditional transaction risk identification methods mainly include risk assessment models based on feature engineering, anomaly detection methods based on statistical laws, and classification models based on machine learning. However, these methods usually only analyze a single stage of the transaction, ignoring the dynamic characteristics of the transaction life cycle, such as the evolution of data features in the transaction initiation, processing and completion stages. In addition, the existing technology lacks an effective comprehensive evaluation mechanism to combine the risk score of the transaction with contextual information, making it impossible to conduct a comprehensive and multi-level risk assessment of multi-stage transaction data. Summary of the invention

[0003] In view of the problems existing in the above-mentioned existing AI-based transaction risk identification methods and devices, the present invention is proposed.

[0004] Therefore, the problem to be solved by the present invention is that the prior art lacks an effective comprehensive assessment mechanism to combine the risk score of a transaction with contextual information, thereby failing to perform a comprehensive and multi-level risk assessment on multi-stage transaction data.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a transaction risk identification method based on AI intelligence, which includes obtaining user transaction information and behavior data in real time and preprocessing the data, extracting data features to generate a feature matrix;

[0006] Generate transaction risk scores based on the feature matrix and perform preliminary risk classification, calculate the positive and negative ideal solution distances of the classified transaction data, and calculate the risk proximity of the transaction based on the ideal solution distance;

[0007] Transaction risks are ranked according to proximity, and risky transactions are managed and controlled based on the ranking.

[0008] As a preferred solution of the AI-based transaction risk identification method of the present invention, the real-time acquisition of user transaction information and behavior data and preprocessing of the data refers to acquiring key fields, user behavior data and context data of each transaction, cleaning and standardizing the collected data, dividing the transactions into peak hours and non-peak hours according to the transaction timestamps of the data, and marking the time context labels;

[0009] The key fields include amount, timestamp, payment method and account information;

[0010] The user behavior data includes the user device ID, login and operation IP address and operation frequency;

[0011] The contextual data includes market fluctuations and geographic events.

[0012] As a preferred solution of the AI-based transaction risk identification method of the present invention, wherein: the extraction of data features to generate a feature matrix refers to extracting basic features, context features, behavior features and cross-features of transaction data based on the time, space and environment information of the transaction;

[0013] The context features include converting the transaction timestamp into periodic hour features and week features, calculating the geographical distance based on the geographical coordinates of the transaction location and the account registration location, and marking the risk level of the transaction environment in combination with hot events;

[0014] The behavior characteristics include calculating the number of transactions and the total transaction amount of the user within the time T, and the number of device switching of the user;

[0015] The cross-feature refers to generating a cross-feature by combining the relationship between different features, including calculating the ratio A of the current transaction amount and the account balance, and calculating the ratio C of the transaction frequency and the account activity;

[0016] Assign a unique transaction ID to each transaction, divide the features into basic feature groups, context feature groups, behavioral feature groups, and cross-feature groups after preprocessing, and store each feature group as a table with the transaction ID as the primary key;

[0017] The preprocessing refers to normalizing the numerical features and encoding the categorical features using Label Encoding;

[0018] The data tables of different feature groups are merged into the main data table according to the transaction ID to obtain the merged feature matrix.

[0019] As a preferred solution of the AI-based transaction risk identification method of the present invention, wherein: generating a transaction risk score based on a feature matrix and performing preliminary risk classification refers to dividing the feature matrix into a basic feature subset, a behavioral feature and a context feature subset, and a cross feature subset according to feature groups;

[0020] Extract basic feature subsets for rapid screening at the transaction initiation stage, use naive Bayes for rapid classification, and calculate the basic feature classification probability of each sample ;

[0021] Set a probability threshold D. If the probability is less than the threshold D, it is a low-risk transaction and is directly approved. Otherwise, it is a high-risk transaction and is marked. A text message verification is sent to the user and the user's single transaction amount is limited. If the user passes the verification, the transaction continues. If not, the transaction is rejected.

[0022] In the transaction processing stage, the behavioral features and context feature subsets of high-risk transactions verified by SMS are extracted to comprehensively construct the dynamic feature matrix X;

[0023] Use random forest to build a nonlinear relationship model between features and risks, perform random forest classification on the dynamic feature matrix X, and output the behavioral feature score ;

[0024] Extract the context feature subset and continue the gradient boosting classification to output the context feature score ;

[0025] Combine the random forest and gradient boosting results to generate the risk score R for the processing stage;

[0026] Set a threshold E. If the risk score is less than the threshold E, it is a medium-risk transaction and requires manual review. Otherwise, it is a high-risk transaction and is marked.

[0027] In the transaction completion stage, the cross-feature subsets of high-risk transactions are extracted, and the historical activity of user accounts of high-risk transactions is calculated to comprehensively construct the final feature matrix;

[0028] Initialize the input layer, hidden layer, and output layer of the multi-layer perceptron model, use historical transaction data to train the model, and calculate the classification error of the current model in each iteration. ;

[0029] Update the weight adjustment factor based on the classification error ;

[0030] Update sample weights using weight adjustment factors;

[0031] Introduce the updated sample weights into the cross loss function ;

[0032] The model parameters are iteratively optimized based on the cross loss function and the Adam optimizer until the loss reaches the minimum and then the iteration is stopped. The iterative model parameters are input into the multi-layer perceptron model to obtain the multi-layer perceptron model.

[0033] The final feature matrix is ​​input into the model input layer, nonlinear features are extracted through the hidden layer, and the probability distribution of each risk category is calculated through the Softmax function in the output layer to obtain the probability of the risk category to which the transaction belongs.

[0034] Select the risk category corresponding to the maximum probability as the classification result of the transaction data;

[0035] The classification results are summarized to form low-risk transaction sets, medium-risk transaction sets, and high-risk transaction sets.

[0036] As a preferred solution of the AI-based transaction risk identification method of the present invention, wherein: the calculation of the positive and negative ideal solution distances of the classified transaction data refers to calculating the feature mean in the low-risk transaction set and generating the feature mean vector as the positive ideal solution, and calculating the feature mean in the high-risk transaction set and generating the feature mean vector as the negative ideal solution;

[0037] Integrate all transaction data features into matrix J and calculate the feature covariance matrix S;

[0038] Use the Mahalanobis distance formula to calculate the distance from each transaction to the positive ideal solution and to the negative ideal solution distance.

[0039] As a preferred solution of the AI-based transaction risk identification method of the present invention, wherein: the risk proximity of the transaction calculated based on the ideal solution distance refers to calculating the transaction risk proximity based on the distance from the calculated transaction data to the positive ideal solution and the distance to the negative ideal solution. ;

[0040] The distance of each transaction from the ideal solution and to the negative ideal solution Substitute the distance into the formula of transaction risk proximity and gradually calculate the .

[0041] As a preferred solution of the AI-based transaction risk identification method described in the present invention, the transaction sorting according to proximity refers to sorting the transaction data in the low-risk transaction set, the medium-risk transaction set and the high-risk transaction set in descending order according to the transaction risk proximity.

[0042] As a preferred solution of the AI-based transaction risk identification method described in the present invention, the management and control of risk transactions based on ranking refers to pushing the risk ranking results to the risk team, and the risk team checks the risk transaction set based on the risk ranking, temporarily freezes the user account related to each transaction in the high-risk transaction, and transfers the transaction amount to the account to be reviewed. For medium-risk transactions and low-risk transactions, they are manually checked in sequence according to the ranking, and the risk transaction set is updated according to the real-time risk value.

[0043] A computer device comprises: a memory and a processor; the memory stores a computer program, characterized in that: when the processor executes the computer program, the steps of a transaction risk identification method based on AI intelligence are implemented.

[0044] 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 a transaction risk identification method based on AI intelligence are implemented.

[0045] The beneficial effects of the present invention are as follows: the present invention significantly improves the accuracy and refinement of risk identification through a multi-stage classification algorithm and positive and negative ideal solution distance calculation, adopts positive and negative ideal solution distance to calculate the risk proximity of transactions, and realizes quantitative sorting of transaction risks. It can not only deeply mine transaction characteristics, but also dynamically optimize the risk assessment model throughout the life cycle of the transaction, significantly improving the efficiency and accuracy of transaction risk identification in complex transaction scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] 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.

[0047] Figure 1 The figure is a flowchart of the transaction risk identification method based on AI intelligence. DETAILED DESCRIPTION

[0048] 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.

[0049] 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.

[0050] 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.

[0051] Example 1, reference Figure 1 , which is the first embodiment of the present invention, and this embodiment provides a transaction risk identification method based on AI intelligence. The transaction risk identification method based on AI intelligence includes the following steps:

[0052] S1. Obtain user transaction information and behavior data in real time and pre-process the data to extract data features and generate a feature matrix;

[0053] Specifically, real-time acquisition of user transaction information and behavior data and preprocessing of the data means obtaining the key fields, user behavior data and context data of each transaction, cleaning and standardizing the collected data, and dividing the transactions into peak hours and off-peak hours according to the transaction timestamps of the data, and marking the time context labels. Through intelligent division of transaction time, the transaction characteristics of peak hours, such as frequent transactions and abnormal capital flow, are captured, and off-peak hours help identify isolated or scattered abnormal transaction behaviors. This time division can enhance the model's ability to identify transaction behavior patterns. By marking the context labels when transactions occur (such as peak hours and off-peak hours), the model's ability to understand the transaction behavior environment is enhanced, especially in the scenario of identifying batch fraud or attacks against peak hours.

[0054] Through real-time data collection, the timeliness of transaction risk assessment is ensured, misjudgment caused by data lag is avoided, and the system's immediate response capability is improved. Real-time collection of behavioral data such as device ID and IP address can also capture dynamic behavioral changes of users, laying the foundation for subsequent risk classification;

[0055] The key fields include amount, timestamp, payment method and account information;

[0056] The user behavior data includes the user device ID, login and operation IP address and operation frequency;

[0057] The contextual data includes market fluctuations and geographic events (promotional activities, regional hot events).

[0058] By acquiring user transaction information, behavior data and contextual data in real time, combined with data cleaning, standardization processing and generation of time context labels, a multi-dimensional and dynamic feature matrix is ​​constructed, providing a comprehensive and real-time analysis basis for transaction risk assessment. Real-time dynamic feature extraction realizes multi-dimensional dynamic monitoring of user behavior and transaction environment, making up for the problem that traditional methods are insufficient in mining transaction timing and contextual features. Context-enhanced analysis introduces external information such as market fluctuations and geographical events, which improves the model's scenario adaptability and environmental sensitivity. Time-sensitive feature labeling enhances the model's ability to identify high-frequency and concentrated transaction risks by dividing peak and non-peak periods.

[0059] Furthermore, extracting data features to generate a feature matrix refers to extracting basic features, context features, behavior features, and cross-features of transaction data based on the time, space, and environment information of the transaction;

[0060] The context features include converting transaction timestamps into periodic hour features and week features, calculating geographic distances based on the geographic coordinates of the transaction location and the account registration location, and marking the risk level of the transaction environment in combination with hot events. For example, major promotional activities are marked as "high transaction volume risk", and transactions in natural disaster areas are marked as "high abnormal possibility";

[0061] By converting timestamps into hourly and weekly features, we can discover the temporal periodic patterns of transaction behaviors (such as the difference in transactions between weekdays and weekends). The calculation of geographical distance can effectively capture abnormal transaction locations (such as a large span between the account registration location and the transaction location). Environmental risk labels generated by hot events (such as natural disasters) provide a key basis for identifying abnormal transactions.

[0062] It realizes hierarchical organization of transaction data and improves analysis efficiency. Different feature groups cover multi-dimensional information such as transaction time, space, and user behavior, providing rich input for risk assessment and supporting dynamic feature expansion (such as adding new hot events or new behavioral indicators) to adapt to changing transaction scenarios.

[0063] The behavior characteristics include calculating the number of transactions and the total transaction amount of the user within the time T, and the number of device switching of the user;

[0064] The quantitative characteristics of transaction behavior can effectively distinguish normal users from potential abnormal users. Behavioral characteristics such as frequent device switching or abnormal high-frequency transactions provide direct clues for identifying fraudulent transactions.

[0065] The cross-feature refers to generating a cross-feature by combining the relationship between different features, including calculating the ratio A of the current transaction amount and the account balance, and calculating the ratio C of the transaction frequency and the account activity;

[0066] Through cross-features (such as the ratio of transaction amount to account balance), hidden transaction behavior patterns are revealed. Cross-features are used as advanced feature inputs to enhance the model's adaptability to complex scenarios.

[0067] Assign a unique transaction ID to each transaction, divide the features into basic feature groups, context feature groups, behavioral feature groups, and cross-feature groups after preprocessing, and store each feature group as a table with the transaction ID as the primary key;

[0068] The preprocessing refers to normalizing numerical features such as amount, account balance, transaction frequency, and geographical distance, and encoding categorical features (such as payment method and market volatility label) using Label Encoding;

[0069] Normalization eliminates unit and magnitude differences between features to ensure the stability of data input. The LabelEncoding method converts categorical features into numerical representations to improve model processing efficiency.

[0070] The data tables of different feature groups are merged into the main data table according to the transaction ID to obtain the merged feature matrix. The feature group is stored with the transaction ID as the primary key to facilitate subsequent data merging and processing. The time, space, behavior and other features of the transaction are integrated into a single matrix to facilitate subsequent analysis. The complete feature matrix provides high-quality input for the machine learning model and improves the classification and evaluation effects.

[0071] By performing refined feature extraction, preprocessing and matrix construction on transaction data, the efficiency and accuracy of transaction risk identification are significantly improved. The introduction of feature classification and context, behavior, and cross-features breaks through the limitations of the single dimension of traditional methods and provides a multi-dimensional analysis perspective for abnormal transactions. By normalizing and encoding standardized data, combined with the unified storage and management of feature matrices, the present invention improves the real-time nature of risk identification, has a high degree of flexibility and adaptability, and can be efficiently applied in a variety of transaction scenarios.

[0072] S2. Generate a transaction risk score based on the feature matrix and perform preliminary risk classification, calculate the positive and negative ideal solution distances of the classified transaction data, and calculate the risk proximity of the transaction based on the ideal solution distance;

[0073] Specifically, generating a transaction risk score and performing preliminary risk classification based on the feature matrix means dividing the feature matrix into a basic feature subset, a behavior feature and context feature subset, and a cross feature subset according to feature groups;

[0074] Extract basic feature subsets for rapid screening at the transaction initiation stage, use naive Bayes for rapid classification, and calculate the basic feature classification probability of each sample :

[0075]

[0076] P

[0077] In the formula, After the basic feature subset V is known, the transaction belongs to the category The probability of and Yes Category and The prior probability of is the basic eigenvalue, Yes Category Lower base features The mean of Yes Category Lower base features The standard deviation of is the eigenvalue In category The conditional probability under is the eigenvalue In category The conditional probability under , n is the total number of features in the basic feature subset, K represents the number of risk classification types, h is the index of the feature, indicating the hth feature currently being processed;

[0078] The probability threshold D is set based on the statistical distribution of historical transaction data. If the probability is less than the threshold D, it is a low-risk transaction and is directly approved. Otherwise, it is a high-risk transaction and is marked. A text message verification is sent to the user and the user's single transaction amount is limited. If the user passes the verification, the transaction continues. If not, the transaction is rejected.

[0079] Naive Bayes is a lightweight algorithm with high computational efficiency. It can quickly complete risk screening at the transaction initiation stage. By directly screening out low-risk transactions, it can reduce the computational pressure in the subsequent stages, provide smooth services for low-risk transactions, implement necessary verification and restrictions on high-risk transactions, and balance security and user convenience.

[0080] In the transaction processing stage, the behavioral features and context feature subsets of high-risk transactions verified by SMS are extracted to comprehensively construct the dynamic feature matrix X;

[0081] Use random forest to build a nonlinear relationship model between features and risks, perform random forest classification on the dynamic feature matrix X, and output the behavioral feature score :

[0082]

[0083] Where T is the total number of decision trees in the random forest, is the classification probability output of the t-th decision tree;

[0084] Extract the context feature subset and continue the gradient boosting classification to output the context feature score :

[0085]

[0086] In the formula, Y is the total number of weak classifiers, which is only used as a counting variable. is the learning rate of the mth weak classifier, is the classification probability output of the mth weak classifier, and Z is the context feature subset;

[0087] Combining the random forest and gradient boosting results generates the risk score R for the processing stage:

[0088]

[0089] In the formula, and are the weights of random forest and gradient boosting models, usually determined by cross-validation and guaranteed to be positive;

[0090] Provide a more balanced risk assessment by weighted fusion of the scoring results of different models;

[0091] Threshold E is set through grid search method. If the risk score is less than threshold E, it is a medium-risk transaction and manual review is required. Otherwise, it is a high-risk transaction and is marked. Dynamic adjustment of threshold can meet the risk management needs in different business scenarios.

[0092] The dynamic feature matrix reflects the changes in trading behavior in real time, improving the real-time and accuracy of classification. The combination of random forest and gradient boosting uses multiple feature data to provide a more comprehensive risk assessment, while considering the risk impact of trading behavior characteristics and environmental characteristics, optimizing the reliability of the scoring results.

[0093] In the transaction completion stage, the cross-feature subsets of high-risk transactions are extracted, and the historical activity of user accounts of high-risk transactions is calculated to comprehensively construct the final feature matrix;

[0094] Initialize the input layer, hidden layer, and output layer of the multi-layer perceptron model, use historical transaction data to train the model, and calculate the classification error of the current model in each iteration. :

[0095]

[0096] In the formula, is the weight of sample i, It is a classification error mark. is the true label of sample i, is the predicted label of sample i, and N is the total number of samples;

[0097] Cross-features enhance the understanding of complex trading behaviors, especially the ability to identify potential risk patterns. Multi-layer perceptrons can efficiently extract nonlinear features from high-dimensional data and achieve more accurate risk classification.

[0098] Update the weight adjustment factor based on the classification error :

[0099]

[0100] Update sample weights using weight adjustment factors :

[0101]

[0102] Introduce the updated sample weights into the cross loss function :

[0103]

[0104] In the formula, is the true label of the kth class of sample i, is the probability that sample i is predicted to be the kth class, K is the total number of risk classification categories;

[0105] Through iterative optimization of the cross loss function and Adam optimizer, the continuous improvement of model performance is ensured;

[0106] The model parameters are iteratively optimized based on the cross loss function and the Adam optimizer until the loss reaches the minimum and then the iteration is stopped. The iterative model parameters are input into the multi-layer perceptron model to obtain the multi-layer perceptron model.

[0107] The final feature matrix is ​​input into the model input layer, nonlinear features are extracted through the hidden layer, and the probability distribution of each risk category is calculated through the Softmax function in the output layer to obtain the probability of the risk category to which the transaction belongs.

[0108] Select the risk category corresponding to the maximum probability as the classification result of the transaction data;

[0109] The classification results are summarized to form low-risk transaction sets, medium-risk transaction sets and high-risk transaction sets. The classification results are combined with the sorting information to provide a basis for priority screening of high-risk transactions, freeze high-risk transactions and transfer them to accounts awaiting review, while low-risk transactions are quickly approved, thereby improving the efficiency and effectiveness of overall risk control, providing clear priorities for manual review, and optimizing resource allocation.

[0110] By constructing dynamic feature matrices and multiple classification models in stages, we have achieved accurate, real-time and intelligent identification of transaction risks. Basic feature screening ensures rapid classification at the transaction initiation stage, the combination of random forest and gradient boosting provides multi-dimensional risk assessment at the processing stage, and the multi-layer perceptron performs in-depth analysis and classification optimization of high-risk transactions at the transaction completion stage.

[0111] Further, calculating the positive and negative ideal solution distances of the classified transaction data refers to calculating the feature mean in the low-risk transaction set and generating the feature mean vector as the positive ideal solution, and calculating the feature mean in the high-risk transaction set and generating the feature mean vector as the negative ideal solution;

[0112] Generating positive and negative ideal solutions based on the feature mean of low-risk transactions and high-risk transactions respectively helps to build a unified reference standard, making transaction risk assessment more targeted and accurate. By extracting the feature mean vector, the noise interference of transaction data is effectively reduced, and the calculation efficiency and result stability are improved;

[0113] Integrate all transaction data features into matrix J and calculate the feature covariance matrix S;

[0114] All transaction features are integrated into a unified matrix form to facilitate batch calculation and analysis, ensuring that the data structure of each transaction is consistent, laying the foundation for covariance matrix and distance calculation. The feature covariance matrix is ​​calculated to consider the correlation between features and eliminate the redundant influence between features, making the distance calculation more accurate, improving the adaptability to high-dimensional feature data, and making the calculation results more robust;

[0115] Use the Mahalanobis distance formula to calculate the distance from each transaction to the positive ideal solution and to the negative ideal solution Distance:

[0116]

[0117]

[0118] In the formula, is the feature vector of the qth transaction, is the positive ideal solution eigenvector, is the negative ideal solution eigenvector, S is the feature covariance matrix, and T is the transposition operation. Compared with the traditional Euclidean distance, the Mahalanobis distance can fully consider the correlation between features, so as to more accurately evaluate the degree of closeness between transaction features and positive and negative ideal solutions, avoid the error problem caused by different feature scales, and enhance the applicability of the model to complex transaction data.

[0119] By introducing the feature covariance matrix S, this method can adapt to the changes in feature distribution of different trading scenarios, ensuring that it still maintains efficient risk identification capabilities in a variety of complex trading environments. Based on the distance calculation of positive and negative ideal solutions and the risk proximity formula, this method can achieve accurate quantification of transaction risks, providing a reliable basis for subsequent sorting and management. The Mahalanobis distance calculation is used to fully consider the correlation between features, avoid result deviations caused by feature scale or noise problems, and improve the stability of the calculation. The feature mean vector and matrix operations are used to realize a batch and structured risk assessment process, which significantly reduces the computational complexity in practical applications.

[0120] S3. Rank transaction risks according to proximity and manage risky transactions based on the ranking;

[0121] Specifically, calculating the risk proximity of a transaction based on the ideal solution distance refers to calculating the risk proximity of a transaction based on the distance of the calculated transaction data to the positive ideal solution and the distance to the negative ideal solution. :

[0122]

[0123] The distance of each transaction from the ideal solution and to the negative ideal solution Substitute the distance into the formula of transaction risk proximity and gradually calculate the .

[0124] By calculating the distance between positive and negative ideal solutions and risk proximity, a multi-dimensional quantitative assessment of transaction risks is achieved, which significantly improves the accuracy of risk assessment. By dynamically adjusting the positive and negative ideal solutions and the characteristic matrix, the present invention can quickly respond to changes in complex transaction environments and ensure that the risk assessment results are always highly consistent with the actual situation. The risk proximity ranking provides a basis for prioritizing high-risk transactions, enabling the risk team to focus limited resources on the most important transactions and optimize the risk management process.

[0125] Furthermore, managing and controlling risky transactions based on ranking means pushing the risk ranking results to the risk team, which checks the risk transaction set based on the risk ranking, temporarily freezes the user account related to each high-risk transaction, and transfers the transaction amount to the account to be reviewed. Medium-risk and low-risk transactions are manually checked in sequence according to the ranking, and the risk transaction set is updated based on the real-time risk value.

[0126] From risk ranking to the management and control of risky transactions, the entire solution realizes the full life cycle management of high-, medium- and low-risk transactions, ensuring that each transaction can be processed in a targeted manner. Through real-time risk value updates and dynamic ranking, the present invention has the ability to respond quickly in complex transaction environments, providing a reliable solution for sudden risk events. The design of freezing high-risk transactions and isolating funds significantly improves the system's ability to prevent and control suspicious transactions, reduces the financial risk exposure during system operation, and allocates processing priorities based on risk ranking, maximizing the utilization of human resources while reducing system resource waste.

[0127] Embodiment 2 is the second embodiment of the present invention. This embodiment is different from the previous embodiment in that:

[0128] 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.

[0129] 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.

[0130] 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.

[0131] 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.

Claims

1. A transaction risk identification method based on AI intelligence, characterized by: include, Obtain user transaction information and behavior data in real time and pre-process the data to extract data features and generate feature matrices; Generate transaction risk scores based on the feature matrix and perform preliminary risk classification, calculate the positive and negative ideal solution distances of the classified transaction data, and calculate the risk proximity of the transaction based on the ideal solution distance; Transaction risks are ranked according to proximity, and risky transactions are managed and controlled based on the ranking.

2. The AI-based transaction risk identification method according to claim 1, characterized in that: The real-time acquisition of user transaction information and behavior data and preprocessing of the data refers to acquiring key fields, user behavior data and context data of each transaction, cleaning and standardizing the collected data, dividing the transactions into peak hours and non-peak hours according to the transaction timestamps of the data, and marking the time context labels; The key fields include amount, timestamp, payment method and account information; The user behavior data includes the user device ID, login and operation IP address and operation frequency; The contextual data includes market fluctuations and geographic events.

3. The AI-based transaction risk identification method according to claim 2, characterized in that: Extracting data features to generate a feature matrix refers to extracting basic features, context features, behavior features, and cross-features of transaction data based on the time, space, and environment information of the transaction; The context features include converting the transaction timestamp into periodic hour features and week features, calculating the geographical distance based on the geographical coordinates of the transaction location and the account registration location, and marking the risk level of the transaction environment in combination with hot events; The behavior characteristics include calculating the number of transactions and the total transaction amount of the user within the time T, and the number of device switching of the user; The cross-feature refers to generating a cross-feature by combining the relationship between different features, including calculating the ratio A of the current transaction amount and the account balance, and calculating the ratio C of the transaction frequency and the account activity; Assign a unique transaction ID to each transaction, divide the features into basic feature groups, context feature groups, behavioral feature groups, and cross-feature groups after preprocessing, and store each feature group as a table with the transaction ID as the primary key; The preprocessing refers to normalizing the numerical features and encoding the categorical features using Label Encoding; The data tables of different feature groups are merged into the main data table according to the transaction ID to obtain the merged feature matrix.

4. The AI-based transaction risk identification method according to claim 3, characterized in that: Generating a transaction risk score based on the feature matrix and performing preliminary risk classification refers to dividing the feature matrix into a basic feature subset, a behavioral feature subset, a context feature subset, and a cross feature subset according to feature groups; Extract basic feature subsets for rapid screening at the transaction initiation stage, use naive Bayes for rapid classification, and calculate the basic feature classification probability of each sample ; Set a probability threshold D. If the probability is less than the threshold D, it is a low-risk transaction and is directly approved. Otherwise, it is a high-risk transaction and is marked. A text message verification is sent to the user and the user's single transaction amount is limited. If the user passes the verification, the transaction continues. If not, the transaction is rejected. In the transaction processing stage, the behavioral features and context feature subsets of high-risk transactions verified by SMS are extracted to comprehensively construct the dynamic feature matrix X; Use random forest to build a nonlinear relationship model between features and risks, perform random forest classification on the dynamic feature matrix X, and output the behavioral feature score ; Extract the context feature subset and continue the gradient boosting classification to output the context feature score ; Combine the random forest and gradient boosting results to generate the risk score R for the processing stage; Set a threshold E. If the risk score is less than the threshold E, it is a medium-risk transaction and requires manual review. Otherwise, it is a high-risk transaction and is marked. In the transaction completion stage, the cross-feature subsets of high-risk transactions are extracted, and the historical activity of user accounts of high-risk transactions is calculated to comprehensively construct the final feature matrix; Initialize the input layer, hidden layer, and output layer of the multi-layer perceptron model, use historical transaction data to train the model, and calculate the classification error of the current model in each iteration. ; Update the weight adjustment factor based on the classification error ; Update sample weights using weight adjustment factors; Introduce the updated sample weights into the cross loss function ; The model parameters are iteratively optimized based on the cross loss function and the Adam optimizer until the loss reaches the minimum and then the iteration is stopped. The iterative model parameters are input into the multi-layer perceptron model to obtain the multi-layer perceptron model. The final feature matrix is ​​input into the model input layer, nonlinear features are extracted through the hidden layer, and the probability distribution of each risk category is calculated through the Softmax function in the output layer to obtain the probability of the risk category to which the transaction belongs. Select the risk category corresponding to the maximum probability as the classification result of the transaction data; The classification results are summarized to form low-risk transaction sets, medium-risk transaction sets and high-risk transaction sets.

5. The AI-based transaction risk identification method according to claim 4, characterized in that: The calculation of the positive and negative ideal solution distances of the classified transaction data refers to calculating the feature mean in the low-risk transaction set and generating the feature mean vector as the positive ideal solution, and calculating the feature mean in the high-risk transaction set and generating the feature mean vector as the negative ideal solution; Integrate all transaction data features into matrix J and calculate the feature covariance matrix S; Use the Mahalanobis distance formula to calculate the distance from each transaction to the positive ideal solution and to the negative ideal solution distance.

6. The AI-based transaction risk identification method according to claim 5, characterized in that: The calculation of the transaction risk proximity based on the ideal solution distance refers to calculating the transaction risk proximity based on the calculated distance of the transaction data to the positive ideal solution and the distance to the negative ideal solution. ; The distance of each transaction from the ideal solution and to the negative ideal solution Substitute the distance into the formula of transaction risk proximity and gradually calculate the .

7. The AI-based transaction risk identification method according to claim 6, characterized in that: The transaction sorting according to proximity refers to sorting the transaction data in the low-risk transaction set, the medium-risk transaction set and the high-risk transaction set in descending order according to the transaction risk proximity.

8. The AI-based transaction risk identification method according to claim 7, characterized in that: The control of risky transactions based on ranking refers to pushing the risk ranking results to the risk team, which checks the risk transaction set based on the risk ranking, temporarily freezes the user account related to each high-risk transaction, transfers the transaction amount to the account to be reviewed, and manually checks medium-risk and low-risk transactions in sequence according to the ranking, and updates the risk transaction set based on the real-time risk value.

9. A computer device comprising: A memory and a processor; the memory stores a computer program, characterized in that: when the processor executes the computer program, the steps of the AI-based transaction risk identification method described in any one of claims 1 to 8 are implemented.

10. 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 transaction risk identification method based on AI intelligence described in any one of claims 1 to 8 are implemented.

Images