Account Risk Level Prediction Method and Device Based on Account Transaction Characteristics

Abstract

This invention provides a method and apparatus for predicting account risk levels based on account transaction features, applicable to the field of artificial intelligence technology. The method includes: extracting and classifying features based on original user information, original account information, and original transaction details to generate account transaction feature matrices for different account categories; training a decision tree based on the account transaction feature matrices for different account categories and pre-generated indicator feature weights to generate account transaction risk prediction models corresponding to different account categories; and using the account transaction risk prediction models corresponding to different account categories to predict the risk of the account transaction features to be identified, generating account risk levels. By combining user information, account information, and transaction details, fine-grained classification of different types of accounts is achieved, improving the comprehensiveness, accuracy, and adaptability of risk prediction, thereby realizing comprehensive, efficient, and accurate anomaly monitoring.

Description

Technical Field

[0001] This invention relates to the field of computer technology, particularly to the field of artificial intelligence technology, and especially to a method and apparatus for predicting account risk levels based on account transaction characteristics. Background Technology

[0002] In today's society, with the continuous advancement of technology, users' methods of disguising abnormal accounts have become increasingly sophisticated and covert. Detecting abnormal transactions is a common challenge in the financial industry. Furthermore, with the continuous development of the internet, the volume of transaction data is growing exponentially, significantly increasing the difficulty for banks and other financial institutions in monitoring abnormal transactions. Traditional methods for detecting abnormal transactions mainly rely on statistical identification based on abnormal transaction rules summarized by experts, such as large single transaction amounts or high transaction frequency within a short period. These methods focus on statistically analyzing account transaction indicators, constructing abnormal transaction rule models, calculating whether transaction indicators are abnormal, and thus determining whether an account is abnormal. Alternatively, machine learning algorithms are used to train models on massive amounts of historical transaction details, and the trained machine learning model is used to determine whether an account is abnormal. However, traditional abnormal transaction rule models rely on pre-set rules for statistical analysis, resulting in slow rule iteration and an inability to adapt to the current situation of rapid economic development and explosive growth in transaction volume. Machine learning algorithms, on the other hand, typically only use account transaction data as a basis, but account anomalies usually involve multiple influencing factors, and there are differences in transactions between different accounts, leading to a high false positive rate and significantly increasing the cost of manual screening. Therefore, both current methods suffer from limitations, poor adaptability, and low accuracy. Summary of the Invention

[0003] One objective of this invention is to provide a method for predicting account risk levels based on account transaction characteristics. This method combines user information, account information, and transaction details to finely segment different types of accounts. A decision tree algorithm is used to generate personalized anomaly identification rules for each account group, predicting the account risk level. This improves the comprehensiveness, accuracy, and adaptability of risk prediction, thereby achieving comprehensive, efficient, and precise anomaly monitoring. Another objective of this invention is to provide a device for predicting account risk levels based on account transaction characteristics. A further objective is to provide a computer-readable medium. A final objective is to provide a computer device.

[0004] To achieve the above objectives, this invention discloses a method for predicting account risk levels based on account transaction characteristics, comprising:

[0005] Obtain original user information, original account information, and original transaction details;

[0006] Based on the original user information, original account information, and original transaction details, feature extraction and classification are performed to generate account transaction feature matrices for different account categories;

[0007] Based on the account transaction feature matrix of different account categories and the pre-generated indicator feature weights, the decision tree is trained to generate account transaction risk prediction models corresponding to different account categories.

[0008] By using account transaction risk prediction models corresponding to different account categories, the risk of the account transaction characteristics to be identified is predicted, and the account risk level is generated.

[0009] Preferably, the original account information includes full account information and abnormal account information;

[0010] Based on the original user information, original account information, and original transaction details, feature extraction and classification are performed to generate account transaction feature matrices for different account categories, including:

[0011] Based on the full account information and the abnormal account information, accounts are associated to generate an account association information table;

[0012] Data cleaning is performed on the original user information, account association information table, and original transaction details information to generate cleaned user information table, account association information table, and transaction details table;

[0013] Data is integrated and accounts are categorized from user information table, account association information table, and transaction details information table to generate account transaction feature matrices for different account categories.

[0014] Preferably, the user information table, account association information table, and transaction details information table are integrated and accounts are classified to generate account transaction feature matrices for different account categories, including:

[0015] The user information table and the account association information table are linked and merged to generate an account feature table;

[0016] Perform one-way data transformation on the transaction details table to generate a one-way transaction table;

[0017] Perform feature transformation on the one-way transaction table to generate a transaction feature table;

[0018] The account feature table and transaction feature table are linked and merged to generate the original account transaction feature table;

[0019] The account feature table is classified using a clustering algorithm to generate account classification results;

[0020] Based on the account classification results and the original account transaction feature table, key features are extracted to generate account transaction feature matrices for different account categories.

[0021] Preferably, the method further includes:

[0022] Key features are extracted from the original account transaction feature table to obtain multiple sets of feature variables and corresponding target variables;

[0023] By using the analytic hierarchy process (AHP), weights are calculated based on multiple sets of characteristic variables and their corresponding target variables to generate indicator feature weights.

[0024] Preferably, the risk prediction model corresponding to different account categories is used to predict the risk of the account transaction characteristics to be identified and generate an account risk level, including:

[0025] Based on the account transaction feature matrix of different account categories, determine the target account category to which the account transaction features to be identified belong;

[0026] Based on the target account classification, the corresponding target account transaction risk prediction model is determined;

[0027] The target account transaction risk prediction model is used to predict the risk of the transaction characteristics of the account to be identified and generate the account risk level.

[0028] Preferably, after predicting the risk of the account transaction characteristics to be identified and generating the account risk level using the account transaction risk prediction model corresponding to different account categories, the method further includes:

[0029] Based on the account risk level, determine whether the current transaction poses any abnormal risks;

[0030] If such a risk control strategy exists, a corresponding risk control solution will be matched based on the account's risk level using a preset risk control strategy.

[0031] This invention also discloses an account risk level prediction device based on account transaction characteristics, comprising:

[0032] The data acquisition unit is used to acquire raw user information, raw account information, and raw transaction details.

[0033] The data processing unit is used to extract and classify features based on the original user information, original account information and original transaction details, and generate account transaction feature matrices for different account categories;

[0034] The model training unit is used to train the decision tree based on the account transaction feature matrix of different account categories and the pre-generated indicator feature weights, and generate account transaction risk prediction models corresponding to different account categories.

[0035] The risk level prediction unit is used to predict the risk of the account transaction characteristics to be identified by using the account transaction risk prediction model corresponding to different account categories, and generate the account risk level.

[0036] The present invention also discloses a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the method described above.

[0037] The present invention also discloses a computer device, including a memory and a processor, wherein the memory is used to store information including program instructions, and the processor is used to control the execution of the program instructions, wherein the processor executes the program to implement the method described above.

[0038] The present invention also discloses a computer program product, including a computer program / instruction, which, when executed by a processor, implements the method described above.

[0039] This invention acquires original user information, original account information, and original transaction details; performs feature extraction and classification based on these information to generate account transaction feature matrices for different account categories; trains a decision tree based on these feature matrices and pre-generated indicator feature weights to generate account transaction risk prediction models corresponding to different account categories; uses these risk prediction models to predict the risk of account transaction features to be identified, generating account risk levels. By combining user information, account information, and transaction details, it performs fine-grained segmentation of different types of accounts, uses a decision tree algorithm to generate personalized anomaly identification rules for each account group, and predicts the risk level of the accounts, thus improving the comprehensiveness, accuracy, and adaptability of risk prediction, thereby achieving comprehensive, efficient, and accurate anomaly monitoring. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 A flowchart illustrating an account risk level prediction method based on account transaction characteristics, provided as an embodiment of the present invention;

[0042] Figure 2 A flowchart illustrating another method for predicting account risk levels based on account transaction characteristics, provided as an embodiment of the present invention;

[0043] Figure 3 A schematic diagram of a data integration structure provided in an embodiment of the present invention;

[0044] Figure 4 A schematic diagram of a feature variable structure based on the analytic hierarchy process provided in an embodiment of the present invention;

[0045] Figure 5 A schematic diagram of the structure of an account risk level prediction device based on account transaction characteristics provided in an embodiment of the present invention;

[0046] Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] It should be noted that the account risk level prediction method and apparatus based on account transaction characteristics disclosed in this application can be used in the field of artificial intelligence technology, or in any field other than artificial intelligence technology. The application field of the account risk level prediction method and apparatus based on account transaction characteristics disclosed in this application is not limited.

[0049] To facilitate understanding of the technical solution provided in this application, the relevant content of the technical solution will be explained below. Regarding the data foundation for account anomaly monitoring, in addition to transaction behavior, a key characteristic of suspicious fund transactions is that the fund transaction contradicts its background, such as the nature of the account, the user's age, and income. Therefore, this invention will combine user information, account information, and transaction data as the data foundation for model training.

[0050] This invention addresses the shortcomings of current abnormal transaction monitoring technologies in the financial industry, which lack analysis of users and their account types and fail to consider the differences in transactions among different account types. It proposes a method for predicting account risk levels that fully integrates user information, account information, and transaction data. This method categorizes different types of accounts with finer granularity, uses a decision tree algorithm to generate personalized anomaly identification rules for each account group, predicts the account's risk level, determines whether transactions are abnormal, and promptly alerts relevant personnel based on the prediction results. This achieves comprehensive, efficient, and accurate anomaly monitoring.

[0051] The following example uses an account risk level prediction device based on account transaction characteristics as the execution subject to illustrate the implementation process of the account risk level prediction method based on account transaction characteristics provided in this embodiment of the invention. It is understood that the execution subject of the account risk level prediction method based on account transaction characteristics provided in this embodiment of the invention includes, but is not limited to, an account risk level prediction device based on account transaction characteristics.

[0052] Figure 1 A flowchart illustrating an account risk level prediction method based on account transaction characteristics provided in this embodiment of the invention is shown below. Figure 1 As shown, the method includes:

[0053] Step 101: Obtain original user information, original account information, and original transaction details.

[0054] Step 102: Extract and classify features based on the original user information, original account information, and original transaction details to generate account transaction feature matrices for different account categories.

[0055] Step 103: Based on the account transaction feature matrix of different account categories and the pre-generated indicator feature weights, train the decision tree to generate account transaction risk prediction models corresponding to different account categories.

[0056] Step 104: Using the account transaction risk prediction model corresponding to different account categories, predict the risk of the account transaction characteristics to be identified and generate the account risk level.

[0057] The technical solution provided in this invention involves obtaining original user information, original account information, and original transaction details; performing feature extraction and classification based on the original user information, original account information, and original transaction details to generate account transaction feature matrices for different account categories; training a decision tree based on the account transaction feature matrices for different account categories and pre-generated indicator feature weights to generate account transaction risk prediction models corresponding to different account categories; using the account transaction risk prediction models corresponding to different account categories to predict the risk of the account transaction features to be identified, generating account risk levels; combining user information, account information, and transaction details to perform fine-grained segmentation of different types of accounts; using a decision tree algorithm to generate personalized anomaly identification rules for each account group; and predicting the risk level of the account, thereby improving the comprehensiveness, accuracy, and adaptability of risk prediction, and achieving comprehensive, efficient, and accurate anomaly monitoring results.

[0058] Figure 2 A flowchart illustrating another account risk level prediction method based on account transaction characteristics provided in this embodiment of the invention is shown below. Figure 2 As shown, the method includes:

[0059] Step 201: Obtain original user information, original account information, and original transaction details.

[0060] In this embodiment of the invention, each step is performed by an account risk level prediction device based on account transaction characteristics.

[0061] In this embodiment of the invention, the distributed database stores the sample dataset required for training the model, and the original user information, original account information, and original transaction details required for training the model are obtained from the distributed database.

[0062] In this embodiment of the invention, the original user information is the basic information of the bank user, including but not limited to the user's gender, age, occupation, annual income, credit rating, phone number, and user identification code; the original account information is the attributes of the account opened by the user for handling fund receipts and payments, including but not limited to the account holder, opening bank, account status, type, card number, account number, and account holder identification code; the original transaction details refer to the transaction records generated by the user using the account for daily receipts and payments, cash deposits and withdrawals, transfers, and other financial activities, including but not limited to the transfer-out account number, transfer-in account number, transaction time, transaction channel, and transaction amount, and the original transaction details can serve as the basis for fund flows.

[0063] Due to the massive amount of original transaction details data, the method involves retrieving original transaction details data within a specified time period. The specified time period can be set according to actual needs, and this embodiment of the invention does not impose any limitations on it. As an optional approach, the specified time period is two years prior to the current time.

[0064] In this embodiment of the invention, the original user information, original account information, and original transaction details are used as the data basis for model training. This approach features a large amount of data and broad coverage, effectively improving the model's precision, accuracy, and comprehensiveness.

[0065] Step 202: Associate accounts based on the full account information and the abnormal account information to generate an account association information table.

[0066] In this embodiment of the invention, the subject of abnormal transaction monitoring is the account. Abnormal transaction monitoring is actually monitoring abnormal accounts and then tracing all transaction links of those accounts. Therefore, the original account information includes full account information and abnormal list account information. The full account information refers to the account information of all accounts in the bank, while the abnormal list account information refers to the account information included in the bank's abnormal list. The abnormal list includes the abnormality level of the abnormal accounts, which includes categories of concern, suspicion, and prohibition.

[0067] Specifically, a new "Abnormal Level" column is added to the full account information as a marker for abnormal accounts. Account numbers in the full account information are matched with account numbers in the abnormal list account information. If a match is found, the corresponding abnormal level for that account number in the abnormal list account information is retrieved, and the retrieved abnormal level value is entered into the corresponding abnormal level in the full account information. If a match fails, the abnormal level value for that account number in the full account information is determined to be normal. The full account information, including the completed "Abnormal Level" column, is defined as the account association information table. The account association information table includes, but is not limited to, account holder, bank, account status, type, card number, account number, and abnormal level.

[0068] Step 203: Clean the original user information, account association information table and original transaction details information to generate cleaned user information table, account association information table and transaction details information table.

[0069] In this embodiment of the invention, the original user information, account association information table, and original transaction details contain some missing, inconsistent, or abnormal data, all of which will affect the training effect of the model. Therefore, data preprocessing is required for the original user information, account association information table, and original transaction details to ensure the rationality and reliability of the subsequent model. Data preprocessing includes data cleaning and data integration. Data cleaning includes, but is not limited to, handling missing values, removing duplicate values, and identifying outliers.

[0070] Specifically, regarding the handling of missing values, if the number of records with missing values ​​in a field is less than a preset record count threshold, the records containing missing values ​​are deleted; if the number of records with missing values ​​in a field is greater than or equal to the preset record count threshold, numerical imputation methods are used to fill in the missing values. Numerical imputation methods include, but are not limited to, mean imputation, mode imputation, fixed value imputation, and interpolation methods.

[0071] It is worth noting that the record count threshold can be set according to actual needs, and this embodiment of the invention does not limit this.

[0072] In this embodiment of the invention, by processing missing values, data resource waste and record omissions can be reduced, data resource utilization can be improved, and model training efficiency can be enhanced.

[0073] Specifically, due to different data storage formats, some data may contain multiple duplicate records. Account association information tables are generally partitioned tables and stored incrementally. For handling duplicate values, the date corresponding to the record with the duplicate value is queried, and the account information for the latest date is used as the account information for that account. For example, if an account modifies an attribute value on day T, the account record will be added to the data table on day T+1. Therefore, if the data table contains attribute information for the same account on both day T and day T+1, the account information for the latest date should be used.

[0074] Specifically, for handling outliers, the first step is to use scatter plots to identify isolated points and outliers in the dataset. These values ​​are considered outliers, and outliers are noise that can interfere with the model training results. If the outlier is erroneous data, it can be corrected or the corresponding record can be deleted. If the outlier is correct data, it can be distributed to different intervals using methods such as binning and clustering to achieve the purpose of handling outliers.

[0075] Step 204: Integrate and classify the user information table, account association information table, and transaction details information table to generate account transaction feature matrices for different account categories.

[0076] In this embodiment of the invention, data integration combines multiple data sources to form a unified dataset, analyzes the relevant characteristics of abnormal transactions from a multi-dimensional perspective, and statistically analyzes the account's inherent characteristics and transaction characteristic indicators, providing a complete data foundation for feature extraction. Specifically, taking the account as the main body, the cleaned user information table, account information table, and transaction details table are integrated to form a complete account transaction characteristic table. Generally, the relationship between users, accounts, and transaction records is one-to-many; a user may have multiple accounts, an account can only belong to one user, an account will correspond to multiple transaction records, and each transaction record will involve multiple accounts. Based on this characteristic, the data is correlated and integrated.

[0077] Figure 3 A schematic diagram of a data integration structure provided in an embodiment of the present invention, such as... Figure 3 As shown, the user information table and the account association information table are linked and merged to obtain the account feature table; the transaction details information table is transformed into a one-way transaction table; the one-way transaction table is transformed into a transaction feature table; and the account feature table and the transaction feature table are linked and merged to obtain the original account transaction feature table.

[0078] In this embodiment of the invention, step 204 specifically includes:

[0079] Step 2041: Merge the user information table and the account association information table to generate an account feature table.

[0080] Specifically, the account holder identification code in the account association information table is matched with the user identification code in the user information table to determine the corresponding user information. The user information of the user to which each account belongs is matched, namely: the basic information of the account holder, and finally an account feature table is generated.

[0081] Step 2042: Perform one-way data transformation on the transaction details information table to generate a one-way transaction table.

[0082] In this embodiment of the invention, the transaction details record includes, but is not limited to, the transfer-out account and the transfer-in account. Since it involves two accounts, it is necessary to take a single account as the main body, add a "transfer-in" and "transfer-out" flag column, split each transaction into two one-way transaction records, and generate a one-way transaction table.

[0083] Furthermore, depending on the transaction method, the flow of funds also differs, exhibiting two phenomena. One is end-to-end direct transactions, such as when user A transfers money to user B; in this case, the funds flow from user A's account to user B's account. The other is indirect transactions, i.e., transactions conducted through third-party payment institutions. For example, when user C shops online and chooses to pay using a third-party application, the funds first flow from user C's account to the third-party payment institution, and then from the third-party payment institution to the shopping platform's account. For indirect transactions, the data table represents two transaction records. These two transaction records need to be merged into one transaction record first. After merging the transactions, the transaction record can be further split to generate a one-way transaction table.

[0084] Step 2043: Perform feature transformation on the one-way transaction table to generate a transaction feature table.

[0085] In this embodiment of the invention, abnormal fund transactions refer to users circulating non-compliant funds in the market through various means, gradually transforming them into funds from legitimate sources. Based on historical experience summarized by experts, some typical characteristics of such transaction accounts include: frequent large-amount fund transactions; short-term dispersed transfers into the account followed by concentrated transfers out, or vice versa; long-dormant accounts suddenly experiencing large-amount fund transactions; frequent cross-regional and cross-bank transactions; and the presence of test transactions, i.e., large-amount transactions following one or two small transactions. Combining these typical characteristics of abnormal transactions, the transaction characteristics of each account are statistically analyzed from dimensions such as time, transaction amount, transaction quantity, and transaction intensity. Transaction characteristics include the number of transactions within a specified time period, total transfer out, total transfer in, average monthly transaction amount, highest transfer in amount, highest transfer out amount, and the number of short-term excessive transactions within a specified time period.

[0086] It is worth noting that the specified time period can be set according to actual needs, and this embodiment of the invention does not limit this. As an optional solution, the specified time period is two years prior to the current time. The number of short-term excess transactions is the number of excess transactions counted monthly.

[0087] Specifically, based on the information in the individual transaction table, statistics and feature transformation are performed according to the transaction characteristics to construct a key indicator system for transactions, thus transforming the individual transaction table into a transaction feature table.

[0088] Step 2044: Link and merge the account feature table and the transaction feature table to generate the original account transaction feature table.

[0089] Specifically, the account numbers in the account feature table are matched with the account numbers in the transaction feature table to determine the corresponding transaction feature information. The corresponding transaction feature information is matched for each account to generate the original account transaction feature table.

[0090] Step 2045: Use a clustering algorithm to classify the account feature table and generate account classification results.

[0091] In this embodiment of the invention, different types of accounts have different transaction characteristics, and the methods for judging anomalies cannot be blindly consistent. For example, some online stores selling digital products and home appliances on shopping platforms have high sales volumes and frequent fund flows, with daily turnovers of hundreds of thousands or even millions of yuan. However, according to the existing identification methods of banking regulatory departments, this situation may be considered an anomaly, leading to misjudgment. Grouping similar accounts into one category can effectively improve the accuracy of the model. Account classification mainly combines the nature of the account itself and the characteristics of the user to which it belongs. Therefore, account classification can be performed based on the account feature table generated in step 2041, removing irrelevant features such as user phone numbers and card numbers, constructing an n×d account feature matrix, where n is the number of accounts and d is the number of feature dimensions. The K-Means clustering algorithm is used to classify the accounts into k categories. The specific calculation method is as follows:

[0092] Step a1: Randomly select k accounts from the n accounts to represent k clusters, where the i-th cluster is defined as... in, M is the d-th feature value of the i-th cluster. i It is the center point of the i-th cluster.

[0093] Step a2: Calculate the probability that each of the other accounts belongs to a specific cluster. The feature vector of the j-th account is defined as X. j =(x j1 ,x j2 ,...,x jd ), j = 1, 2, ..., n, where xjd Let d be the feature value of the j-th account, and the Euclidean distance between the j-th account and the center of the i-th cluster is:

[0094]

[0095] Where, x jd Let r be the d-th feature value of the j-th account. ji M is the Euclidean distance between the j-th account and the i-th cluster center. i Let X be the center point of the i-th cluster. j Let be the feature vector of the j-th account.

[0096] The probability that the j-th account belongs to the i-th cluster can be defined as:

[0097]

[0098] Where, r ji p is the Euclidean distance between the j-th account and the i-th cluster center. ji Let be the probability that the j-th account belongs to the i-th cluster, and k be the number of clusters.

[0099] According to the principle of the K-Means clustering algorithm, the smaller the distance between the account feature vector and the cluster center, the higher the similarity and the greater the probability of being assigned to that cluster. That is, the probability of a sample point belonging to a cluster is inversely proportional to its Euclidean distance, and each account is assigned to the cluster with the highest probability.

[0100] Step a3: Update the cluster centers. Recalculate the center point of each cluster to minimize the sum of squared errors (SSE). Repeat step a2 until the cluster center points no longer change or the maximum number of iterations is reached. Output the account classification results.

[0101]

[0102] Where SSE is the sum of squared errors, r ji Let k be the Euclidean distance between the j-th account and the center of the i-th cluster, and k be the number of clusters.

[0103] Step 2046: Extract key features based on the account classification results and the original account transaction feature table to generate account transaction feature matrices for different account categories.

[0104] In this embodiment of the invention, key feature extraction involves filtering data by indicators and mining key features for training the abnormal transaction monitoring model. The dataset obtained from data preprocessing contains many features that are not very relevant to the classification and identification of abnormal transactions or are even irrelevant. For example, features such as users' phone numbers and card numbers may not be relevant to the target of this model and need to be removed from the original feature set. Key feature extraction reduces the impact of noise on the model.

[0105] Specifically, based on the account classification results, key features are extracted from the original account transaction feature tables corresponding to different account classifications to generate an account transaction feature matrix for each account classification.

[0106] Step 205: Extract key features from the original account transaction feature table to obtain multiple sets of feature variables and corresponding target variables.

[0107] In this embodiment of the invention, features irrelevant to the model objective, such as the user's phone number and card number, are removed. The impact of noise on the model is reduced by extracting key features, and the extracted account transaction features are obtained.

[0108] In this embodiment of the invention, the training set of the supervised machine learning classification algorithm should include feature variables and target variables for each sample. Feature variables are the characteristic attributes of the account, i.e., the attributes of the account transaction feature table. Obvious irrelevant attributes in the account transaction feature table are removed, such as user phone numbers, card numbers, and user identification codes, to reduce data noise. The processed feature variables include: user characteristics (age, occupation, annual income, credit rating, etc.), account characteristics (account status, account nature, account category, account opening duration, account holder identification code, etc.), and transaction behavior (number of transactions within a specified time period, total amount transferred out, total amount transferred in, average monthly transaction amount, highest amount transferred in, highest amount transferred out, number of times short-term excessive transactions occur within a specified time period, etc.). The target variable is the target classification value for the account, i.e., the abnormality level in the account characteristics. Banks generally classify account abnormality levels into four categories: normal, watchlist, suspicious, and prohibited. This data classification has already undergone preliminary screening and positioning by the bank and has a certain degree of reliability, making it suitable as the target variable.

[0109] Step 206: Using the Analytic Hierarchy Process (AHP), calculate the weights of multiple sets of feature variables and their corresponding target variables to generate indicator feature weights.

[0110] In this embodiment of the invention, the characteristic variables of the sample are combined with the target variable, and the initial weight parameters of the index are adjusted by introducing the idea of ​​the Analytic Hierarchy Process (AHP). Subjective judgment is combined with objective analysis, and the initial weights that conform to the logic of reality are obtained by using systematic, comparative and comprehensive thinking. Figure 4 A schematic diagram of a feature variable structure based on the analytic hierarchy process (AHP) provided in an embodiment of the present invention is shown below. Figure 4 As shown, the feature variable structure includes a target layer, a criterion layer, and an indicator layer. The target layer is the judgment target variable, namely: identifying abnormal accounts (N); the criterion layer is the feature category, namely: user characteristics (C1), account characteristics (C2), and transaction characteristics (C3); the indicator layer is the secondary indicator under the criterion layer, namely: the aforementioned feature variables, namely, age (P11), occupation (P12), ..., credit rating (P1r), account status (P21), account type (P22), ..., account opening duration (P2s), number of transactions (P31), total amount transferred out (P32), ..., number of excessive transactions (P3t). Each layer constructs a judgment matrix by comparing the pairwise importance of the indicators, and then calculates the weight of the secondary indicators, i.e., the feature variables.

[0111] By comparing the importance of criterion layers C1, C2, and C3 with respect to target layer N pairwise, the judgment matrix between the target layer and the criterion layer is obtained. Among them, a ij a is the comparison factor between Ci and Cj. ij a ji =1.

[0112] The weights of each factor in the criterion layer are calculated using the arithmetic mean method. The weight vector of the criterion layer is obtained as u = (u1, u2, u3). Where u i a represents the weight of feature i in the criterion layer. ij is the comparison factor between Ci and Cj.

[0113] For each indicator layer P, a pairwise comparison is performed to determine its importance relative to its respective parent criterion layers C1, C2, and C3, resulting in the following judgment matrices: Where k = 1, 2, 3, corresponding to the three factors of the criterion layer, b ij (k) b is the comparison factor for Pki and Pkj under the criterion layer Ck. ij (k) b ji (k) =1.

[0114] The index weights under each criterion layer Ck are calculated using the arithmetic mean method. Where i = 1, 2, ..., m, m represents the number of indicators under the criterion layer Ck, and the weight vector of the indicator layer under the criterion layer Ck is v. (k) =(v1) (k) v2 (k) ,…,v m (k) ), where v i (k) Let a be the weight of index i under criterion layer Ck. ijis the comparison factor between Ci and Cj.

[0115] By combining the corresponding weights of the criterion layer and the indicator layer, the indicator feature weights of the indicator layer are finally calculated:

[0116] w=(u1v (1) u2v (2) u3v (3) )=(u1v1 (1) ,…,u1v r (r) u2v s (2) ,…,u2v s (s) u3v t (3) ,…,u3v t (t) )

[0117] Where w represents the initial indicator feature weights for subsequent model use; u1, u2, and u3 are the user feature weights, account feature weights, and transaction feature weights, respectively; v (1) v (2) v (3) These are the indicator weights for each indicator feature under user characteristics, account characteristics, and transaction characteristics, respectively.

[0118] Step 207: Train the decision tree based on the account transaction feature matrix of different account categories and the pre-generated indicator feature weights to generate account transaction risk prediction models corresponding to different account categories.

[0119] In this embodiment of the invention, the account transaction feature matrix based on different account categories is used as the training set. Combined with the indicator feature weights, the indicator bias problem of the model training algorithm is corrected. The decision tree algorithm is used to train the model on the data in the account transaction feature matrix of different account categories to generate personalized account transaction risk prediction models corresponding to different account categories.

[0120] Decision tree classification is a model that infers tree-shaped classification rules from an unordered and unregulated training sample set. A decision tree contains a root node, several internal nodes, and several leaf nodes. The root node and internal nodes store the result splits for each feature variable, while the leaf nodes correspond to the target variable. Each branch path from the root node to a leaf node can form a rule sequence. Since the feature variables in the training sample set include both discrete features (such as credit rating, account type, etc.) and continuous variables (such as age, transaction amount, etc.), the model is trained based on the C4.5 algorithm. This algorithm uses the information gain ratio of the feature variables to split the decision tree, prioritizing features with higher information gain ratios as splitting nodes, gradually reducing the system's clutter until the optimal classification decision tree is obtained.

[0121] Specifically, decision tree models are trained for different account categories to generate personalized anomaly detection rules for each category. Taking account category A as an example, there are N... a Given a sample set of accounts of type a, calculate the information entropy H(a) of the target variable. This is used to measure the degree of system disorder in the sample set of class A accounts, where k is the number of categories of the target variable. That is, as mentioned above, abnormal account levels are divided into four categories: normal, monitored, suspicious, and prohibited. Therefore, k = 4, D i This represents the number of samples divided into each target variable value.

[0122] Calculate feature x in class A account d Its own information entropy H(x) d ), The C4.5 algorithm calculates the information entropy directly from the feature values ​​for discrete features. For continuous features, it sorts the values ​​in ascending order, takes the midpoint between two adjacent values ​​as the split point, discretizes the continuous feature, and then calculates the information entropy. Here, m is the number of values ​​for the feature variable, and D... v The number of samples for each value of the feature variable.

[0123] The essence of the decision tree algorithm is to find the best split for each feature and the order in which different features are split, thereby obtaining a decision tree with optimal classification performance. This is based on the feature variable x. d After partitioning, calculate the overall information entropy of the samples, i.e., the conditional entropy H(a,x). d ), This leads to the information gain Gain(a,x) of the feature segmentation results. d Gain(a,x) d )=H(a)-H(a,x d The iteration makes the Gain(a,x) of each feature equal to the sum of its values. dTo maximize the optimal partition within each feature, we find the best partition within each feature. Here, m is the number of values ​​for the feature variable, and D... v N is the number of samples for each value of the feature variable, k is the number of categories of the target variable, and N is the number of samples for each value of the feature variable. a This represents the number of samples.

[0124] After finding the optimal split for each feature variable, it's necessary to prioritize one of the feature variables as the first split node in the decision tree. Generally, information gain is affected by the number of values ​​a feature can take and is not suitable for comparing features. The fewer values ​​a feature variable has, the smaller its conditional entropy and the greater its information gain. If we also use information gain to determine feature priority, the model will favor the feature variable with fewer values. Therefore, C4.5 uses information gain ratio. To enhance the comparability between features and prevent the model from biasing towards features with fewer values, the split of the feature variable with the highest information gain ratio is used as the first split node of the decision tree. While information gain ratio corrects for feature bias, it primarily relies on the system's internal data structure. However, abnormal transaction monitoring scenarios involve user, account, and transaction features, exhibiting considerable complexity and flexibility. Therefore, it's crucial to incorporate real-world experience and consider the inherent properties of the features themselves to improve the model's objectivity. This ensures the model accurately and completely reflects the actual situation. Furthermore, by combining indicator feature weights and optimizing the information gain ratio, the model can prioritize features with higher importance when they have similar or identical information gain ratios.

[0125]

[0126] Where Gain_rate′(a,x) d ) represents the optimized information gain ratio, w d For feature x d The indicator feature weights, H(a) is the information entropy of the target variable of the sample set of account class a, H(a,x) d H(x) is the conditional entropy. d ) is a feature x d Its own information entropy.

[0127] After selecting the optimized feature with high information gain ratio as the first splitting node, the sample data is divided into multiple sub-sample sets according to the feature. A new splitting node is selected for each sub-sample set according to the above algorithm, and so on, until each sub-sample set belongs to the same target variable and no further splitting is needed. Finally, an optimal decision tree is obtained, and the branch path from the root node to the leaf node can be extracted as an abnormal account monitoring rule.

[0128] Step 208: Based on the account transaction feature matrix of different account categories, determine the target account category of the account transaction features to be identified.

[0129] In this embodiment of the invention, based on the transaction characteristics of the account to be identified, the Euclidean distance is calculated with the center point of the cluster of each account category; multiple Euclidean distances are compared to determine the minimum Euclidean distance; and the account category corresponding to the minimum Euclidean distance is determined as the target account category.

[0130] Step 209: Based on the target account classification, determine the corresponding target account transaction risk prediction model.

[0131] In this embodiment of the invention, a personalized account transaction risk prediction model for the corresponding cluster is selected based on the target account classification; the selected account transaction risk prediction model is then determined as the target account transaction risk prediction model.

[0132] Step 210: Using the target account transaction risk prediction model, predict the risk of the transaction characteristics of the account to be identified and generate the account risk level.

[0133] Specifically, the transaction characteristics of the account to be identified are input into the target account transaction risk prediction model to predict the account risk level, and the account risk level is output.

[0134] Step 211: Based on the account risk level, determine whether there is any abnormal risk in the current transaction. If there is, proceed to step 212; otherwise, the process ends.

[0135] In this embodiment of the invention, the account risk level includes normal, watch, suspicious, and prohibited categories.

[0136] Specifically, if the account risk level is suspicious or prohibited, it indicates that there is an abnormal risk in the current transaction, and step 212 is executed; if the account risk level is normal or watchful, it indicates that there is no abnormal risk in the current transaction, and the process ends.

[0137] Step 212: Based on the account risk level, match the corresponding risk control solution according to the preset risk control strategy.

[0138] In this embodiment of the invention, the risk control strategy can be set according to actual needs, and this embodiment of the invention does not limit it. As an optional solution, if the account risk level is suspicious, the corresponding risk control solution is for relevant bank staff to follow up on the transaction chain of the current transaction funds, quickly locate the account transaction chain through the one-way transaction table, and conduct risk control analysis; if the prediction result is prohibited, the corresponding risk control solution is for the bank system to immediately intercept the flow of funds and freeze the relevant account.

[0139] This invention addresses the current financial industry's need for efficient and accurate transaction anomaly monitoring. It comprehensively analyzes user characteristics, account characteristics, and transaction characteristics to uncover indicators influencing transaction risk and extract features to reduce noise. The analytic hierarchy process (AHP) is introduced to calculate the weights of transaction features for each account, and expert evaluation is used to subjectively quantify the importance of features. Considering the differences in transaction patterns among different account types, K-Means clustering is employed to further refine the segmentation of all accounts. An improved decision tree algorithm is used to train personalized models for different account groups, resulting in a unique decision tree model for each account group. This extracts differentiated anomaly monitoring rules, effectively improving prediction accuracy. The decision tree algorithm incorporates feature weights to adjust the information gain ratio, eliminating subjective bias and objective bias, thereby selecting the optimal feature segmentation to construct the decision tree and enhancing model reliability.

[0140] It is worth noting that the acquisition, storage, use, and processing of data in the technical solution of this application all comply with relevant laws and regulations. The user information in the embodiments of this application was obtained through legal and compliant means, and the acquisition, storage, use, and processing of user information have been authorized and agreed upon by the users.

[0141] It is worth noting that the information collected in this application is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data all comply with the relevant laws, regulations and standards of the relevant countries and regions, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding operation portals are provided for users to choose to authorize or refuse.

[0142] It is worth noting that the technical solution provided in this application provides users with a corresponding operation entry point, allowing users to choose to agree to or reject the automated decision-making result; if the user chooses to reject, the process will proceed to the expert decision-making process.

[0143] The technical solution of the account risk level prediction method based on account transaction features provided in this invention involves obtaining original user information, original account information, and original transaction details; performing feature extraction and classification based on the original user information, original account information, and original transaction details to generate account transaction feature matrices for different account categories; training a decision tree based on the account transaction feature matrices for different account categories and pre-generated indicator feature weights to generate account transaction risk prediction models corresponding to different account categories; using the account transaction risk prediction models corresponding to different account categories to predict the risk of the account transaction features to be identified, generating account risk levels. By combining user information, account information, and transaction details, fine-grained classification of different types of accounts is performed, and a decision tree algorithm is used to generate personalized anomaly identification rules for each type of account group to predict the risk level of the account, thereby improving the comprehensiveness, accuracy, and adaptability of risk prediction, and achieving comprehensive, efficient, and accurate anomaly monitoring results.

[0144] Figure 5 This is a schematic diagram of an account risk level prediction device based on account transaction characteristics provided in an embodiment of the present invention. This device is used to execute the aforementioned account risk level prediction method based on account transaction characteristics, such as... Figure 5 As shown, the device includes: a data acquisition unit 11, a data processing unit 12, a model training unit 13, and a risk level prediction unit 14.

[0145] The data acquisition unit 11 is used to acquire original user information, original account information and original transaction details.

[0146] The data processing unit 12 is used to extract and classify features based on the original user information, original account information and original transaction details, and generate account transaction feature matrices for different account categories.

[0147] The model training unit 13 is used to train the decision tree based on the account transaction feature matrix of different account categories and the pre-generated indicator feature weights, so as to generate the account transaction risk prediction model corresponding to different account categories.

[0148] The risk level prediction unit 14 is used to predict the risk of the account transaction characteristics to be identified by using the account transaction risk prediction model corresponding to different account categories, and generate the account risk level.

[0149] In this embodiment of the invention, the original account information includes full account information and abnormal list account information; the data processing unit 12 is specifically used to associate accounts based on the full account information and abnormal list account information to generate an account association information table; to clean the original user information, account association information table and original transaction details information to generate cleaned user information table, account association information table and transaction details information table; and to integrate and classify the user information table, account association information table and transaction details information table to generate account transaction feature matrices for different account categories.

[0150] In this embodiment of the invention, the data processing unit 12 is specifically used to: merge the user information table and the account association information table to generate an account feature table; perform one-way data transformation on the transaction detail information table to generate a one-way transaction table; perform feature transformation on the one-way transaction table to generate a transaction feature table; merge the account feature table and the transaction feature table to generate an original account transaction feature table; classify the accounts in the account feature table using a clustering algorithm to generate account classification results; and extract key features based on the account classification results and the original account transaction feature table to generate account transaction feature matrices for different account categories.

[0151] In this embodiment of the invention, the device further includes a feature extraction unit 15 and a feature weight generation unit 16.

[0152] The feature extraction unit 15 is used to extract key features from the original account transaction feature table to obtain multiple sets of feature variables and corresponding target variables.

[0153] The feature weight generation unit 16 is used to generate indicator feature weights by performing weight calculations based on multiple sets of feature variables and corresponding target variables using the analytic hierarchy process.

[0154] In this embodiment of the invention, the risk level prediction unit 14 is specifically used to determine the target account category to which the account transaction feature to be identified belongs based on the account transaction feature matrix of different account categories; determine the corresponding target account transaction risk prediction model based on the target account category; and perform risk prediction on the account transaction feature to be identified through the target account transaction risk prediction model to generate an account risk level.

[0155] In this embodiment of the invention, the device further includes: an abnormal risk judgment unit 17 and a risk control scheme matching unit 18.

[0156] The abnormal risk assessment unit 17 is used to determine whether there is any abnormal risk in the current transaction based on the account risk level.

[0157] The risk control scheme matching unit 18 is used to match the corresponding risk control scheme according to the account risk level based on the preset risk control strategy if it exists.

[0158] In this embodiment of the invention, original user information, original account information, and original transaction details are obtained; features are extracted and classified based on the original user information, original account information, and original transaction details to generate account transaction feature matrices for different account categories; a decision tree is trained based on the account transaction feature matrices for different account categories and pre-generated indicator feature weights to generate account transaction risk prediction models corresponding to different account categories; through the account transaction risk prediction models corresponding to different account categories, risk prediction is performed on the account transaction features to be identified, and account risk levels are generated. By combining user information, account information, and transaction details, fine-grained classification of different types of accounts is performed, and a decision tree algorithm is used to generate personalized anomaly identification rules for each type of account group to predict the risk level of the account, thereby improving the comprehensiveness, accuracy, and adaptability of risk prediction, and thus achieving a comprehensive, efficient, and accurate anomaly monitoring effect.

[0159] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer device, specifically, a computer device can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any combination of these devices.

[0160] This invention provides a computer device, including a memory and a processor. The memory is used to store information including program instructions, and the processor is used to control the execution of the program instructions. When the program instructions are loaded and executed by the processor, they implement the steps of the above-described embodiment of the account risk level prediction method based on account transaction characteristics. For a detailed description, please refer to the above-described embodiment of the account risk level prediction method based on account transaction characteristics.

[0161] The following is for reference. Figure 6 It shows a schematic diagram of the structure of a computer device 600 suitable for implementing the embodiments of this application.

[0162] like Figure 6 As shown, the computer device 600 includes a central processing unit (CPU) 601, which can perform various appropriate tasks and processes based on programs stored in read-only memory (ROM) 602 or programs loaded from storage section 608 into random access memory (RAM) 603. The RAM 603 also stores various programs and data required for the operation of the computer device 600. The CPU 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0163] The following components are connected to I / O interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal feedback (LCD), etc., and speakers, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to I / O interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 610 as needed so that computer programs read from it can be installed in storage section 608 as needed.

[0164] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program including program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 609, and / or installed from removable medium 611.

[0165] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0166] For ease of description, the above devices are described separately by function as various units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0167] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0168] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0169] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0170] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0171] The acquisition, storage, use, and processing of data in this application all comply with the relevant provisions of national laws and regulations.

[0172] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0173] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0174] This application can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0175] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0176] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for predicting account risk level based on account transaction characteristics, characterized in that, The method includes: Obtain original user information, original account information, and original transaction details, wherein the original account information includes full account information and abnormal account list information; Based on the original user information, original account information, and original transaction details, feature extraction and classification are performed to generate account transaction feature matrices for different account categories. This includes: associating accounts based on the full account information and abnormal account information to generate an account association information table; cleaning the original user information, account association information table, and original transaction details to generate cleaned user information table, account association information table, and transaction details table; integrating and classifying the user information table, account association information table, and transaction details table to generate account transaction feature matrices for different account categories, including merging the user information table and account association information table to generate an account feature table; performing one-way data transformation on the transaction details table to generate a one-way transaction table; performing feature transformation on the one-way transaction table to generate a transaction feature table; merging the account feature table and transaction feature table to generate an original account transaction feature table; classifying accounts in the account feature table using a clustering algorithm to generate account classification results; and extracting key features based on the account classification results and the original account transaction feature table to generate account transaction feature matrices for different account categories. Based on the account transaction feature matrix of the different account categories and the pre-generated indicator feature weights, the decision tree is trained to generate account transaction risk prediction models corresponding to different account categories. Using the account transaction risk prediction models corresponding to the different account categories, the risk prediction of the account transaction characteristics to be identified is performed to generate an account risk level. Specifically, this includes: determining the target account category to which the account transaction characteristics to be identified belong based on the account transaction characteristic matrix of the different account categories; determining the corresponding target account transaction risk prediction model based on the target account category; and using the target account transaction risk prediction model to predict the risk of the account transaction characteristics to be identified to generate an account risk level.

2. The account risk level prediction method based on account transaction characteristics according to claim 1, characterized in that, The method further includes: Key features are extracted from the original account transaction feature table to obtain multiple sets of feature variables and corresponding target variables; The indicator feature weights are generated by calculating the weights of multiple sets of feature variables and their corresponding target variables using the analytic hierarchy process.

3. The account risk level prediction method based on account transaction characteristics according to claim 1, characterized in that, After generating account risk levels by predicting the risk of the account transaction characteristics to be identified using the account transaction risk prediction model corresponding to the different account categories, the method further includes: Based on the account risk level, determine whether the current transaction carries any abnormal risks; If such a risk control strategy exists, a corresponding risk control solution will be matched based on the account's risk level using a preset risk control strategy.

4. An account risk level prediction device based on account transaction characteristics, used to execute the account risk level prediction method based on account transaction characteristics as described in claim 1, characterized in that, The device includes: The data acquisition unit is used to acquire raw user information, raw account information, and raw transaction details. The data processing unit is used to extract and classify features based on the original user information, original account information and original transaction details, and generate account transaction feature matrices for different account categories; The model training unit is used to train the decision tree based on the account transaction feature matrix of the different account categories and the pre-generated indicator feature weights, and generate account transaction risk prediction models corresponding to different account categories. The risk level prediction unit is used to predict the risk of the account transaction characteristics to be identified by using the account transaction risk prediction model corresponding to the different account categories, and generate the account risk level.

5. A computer-readable medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the account risk level prediction method based on account transaction characteristics as described in any one of claims 1 to 3.

6. A computer device comprising a memory and a processor, the memory for storing information including program instructions, and the processor for controlling the execution of the program instructions, characterized in that, When the program instructions are loaded and executed by the processor, they implement the account risk level prediction method based on account transaction characteristics as described in any one of claims 1 to 3.

7. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it implements the account risk level prediction method based on account transaction characteristics as described in any one of claims 1 to 3.

Images