Transaction reporting generation method and system for money laundering monitoring

Abstract

This invention relates to a method and system for generating transaction reports for anti-money laundering monitoring, relating to the field of data monitoring. The method includes: monitoring transactions of a target account; calculating and obtaining short-term transaction amounts and frequencies when fluctuating transaction amounts occur, and performing amount anomaly analysis and frequency anomaly analysis to obtain target amount anomaly coefficients and target frequency anomaly coefficients; obtaining a related amount network and a related frequency network; predicting the split transaction compensation ratio and split quantity based on short-term transaction amounts to obtain the split transaction compensation ratio and split quantity; performing related transaction anomaly analysis using the split transaction compensation ratio and split quantity to obtain related amount anomaly coefficients and related frequency anomaly coefficients, and generating a transaction report. This invention solves the technical problems of inaccurate, incomplete, and false alarm information in existing anti-money laundering transaction monitoring reports.

Description

Technical Field

[0001] This invention relates to the field of data monitoring technology, and in particular to a method and system for generating transaction reports for anti-money laundering monitoring. Background Technology

[0002] Currently, money laundering methods include splitting transactions to conceal the source of illicit funds, making it difficult for traditional anti-money laundering monitoring systems to accurately identify suspicious transactions. Splitting transactions, by mixing large sums of money with funds from other sources and dividing them into multiple smaller transactions, can disguise and obscure the source of funds, making it difficult to regulate.

[0003] Current methods for monitoring suspicious money laundering transactions mostly rely on simple judgment rules, such as determining whether the amount is excessive, and generating corresponding transaction monitoring reports. However, it is currently difficult to accurately monitor and identify the characteristics of split transactions, leading to technical problems such as inaccurate, incomplete, and false alarm information in the transaction monitoring reports. Summary of the Invention

[0004] This invention addresses the technical problems of inaccurate, incomplete, and false alarm information in existing transaction monitoring reports.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] In a first aspect, the present invention provides a method for generating transaction reports for anti-money laundering monitoring, comprising: monitoring transactions of a target account; when fluctuating transaction amounts occur, calculating and obtaining short-term transaction amounts and short-term transaction frequencies; and performing amount anomaly analysis and frequency anomaly analysis to obtain target amount anomaly coefficients and target frequency anomaly coefficients.

[0007] Obtain the network of associated accounts of the target account, monitor and calculate the associated short-term transaction amount and associated short-term transaction frequency of multiple associated accounts within the network of associated accounts, and obtain the associated amount network and associated frequency network;

[0008] Based on the short-term transaction amount, the compensation ratio and number of split transactions are predicted to obtain the compensation ratio and number of split transactions.

[0009] Based on the associated amount network and associated frequency network, the associated transaction anomaly analysis is performed using the split transaction compensation ratio and split quantity to obtain the associated amount anomaly coefficient and associated frequency anomaly coefficient. Combined with the target amount anomaly coefficient and target frequency anomaly coefficient, the total anomaly coefficient is calculated, and a transaction report is generated.

[0010] The transaction monitoring of the target account is carried out. When the transaction amount fluctuates, the short-term transaction amount and short-term transaction frequency are calculated and obtained. The amount anomaly analysis and frequency anomaly analysis are performed to obtain the target amount anomaly coefficient and the target frequency anomaly coefficient.

[0011] Obtain the network of associated accounts of the target account, monitor and calculate the associated short-term transaction amount and associated short-term transaction frequency of multiple associated accounts within the network of associated accounts, and obtain the associated amount network and associated frequency network;

[0012] Based on the short-term transaction amount, the compensation ratio and number of split transactions are predicted to obtain the compensation ratio and number of split transactions.

[0013] Based on the associated amount network and associated frequency network, the associated transaction anomaly analysis is performed using the split transaction compensation ratio and split quantity to obtain the associated amount anomaly coefficient and associated frequency anomaly coefficient. Combined with the target amount anomaly coefficient and target frequency anomaly coefficient, the total anomaly coefficient is calculated, and a transaction report is generated.

[0014] Secondly, the present invention provides a transaction report generation system for anti-money laundering monitoring, comprising:

[0015] The target anomaly analysis module is used to monitor the transactions of target accounts. When fluctuations occur in transaction amounts, it calculates and obtains short-term transaction amounts and short-term transaction frequencies, and performs amount anomaly analysis and frequency anomaly analysis to obtain the target amount anomaly coefficient and the target frequency anomaly coefficient.

[0016] The associated network acquisition module is used to acquire the associated account network of the target account, monitor and calculate the associated short-term transaction amount and associated short-term transaction frequency of multiple associated accounts within the associated account network, and obtain the associated amount network and associated frequency network.

[0017] The splitting feature prediction module is used to predict the splitting transaction compensation ratio and the number of splits based on the short-term transaction amount, and to obtain the splitting transaction compensation ratio and the number of splits.

[0018] The report generation module is used to perform anomaly analysis of related transactions based on the related amount network and the related frequency network, using the split transaction compensation ratio and the number of splits, to obtain the related amount anomaly coefficient and the related frequency anomaly coefficient. Combining the target amount anomaly coefficient and the target frequency anomaly coefficient, the module calculates the total anomaly coefficient and generates a transaction report.

[0019] The beneficial effects of this invention are as follows: By conducting in-depth analysis of the transaction behavior of the target account, this invention improves the ability to identify split transactions, thereby effectively reducing the problems of inaccurate, incomplete, and false alarm problems in traditional transaction monitoring reports. Firstly, when monitoring the transactions of the target account, by calculating short-term transaction amounts and frequencies, and performing amount anomaly analysis and frequency anomaly analysis, the target amount anomaly coefficient and target frequency anomaly coefficient are obtained. This allows for accurate identification of short-term transaction fluctuations, which is more flexible and adaptable compared to traditional methods that rely solely on fixed amount thresholds. Secondly, by acquiring the network of related accounts of the target account and monitoring and calculating the related short-term transaction amounts and frequencies of multiple related accounts, a related amount network and a related frequency network are formed. This technical feature can effectively identify multiple related accounts involved in the split transaction process, avoiding monitoring blind spots caused by analyzing only a single account, and is applicable to the characteristics of split transactions. Furthermore, by predicting the split transaction compensation ratio and split quantity based on the short-term transaction amount, it is possible to effectively predict changes in funds and transaction frequency caused by split transactions, making the monitoring and analysis more accurate. Subsequently, anomaly analysis of related transactions is conducted using related amount networks and related frequency networks, combined with the compensation ratio and number of split transactions. This yields anomaly coefficients for related amount and frequency, quantifying the suspiciousness of split transactions and improving the accuracy of anomaly identification. Finally, the total anomaly coefficient is calculated by combining the target amount and target frequency anomaly coefficients. A transaction report is then generated based on short-term transaction amounts, short-term transaction frequencies, and related account networks, ensuring more comprehensive report content and enhancing the ability to accurately monitor money laundering activities related to split transactions, thereby improving the accuracy and effectiveness of anti-money laundering monitoring. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the transaction report generation method for anti-money laundering monitoring provided by the present invention.

[0021] Figure 2 This is a schematic diagram of the transaction report generation system for anti-money laundering monitoring provided by the present invention.

[0022] Figure labeling: Target anomaly analysis module 11, correlation network acquisition module 12, split feature prediction module 13, report generation module 14. Detailed Implementation

[0023] 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.

[0024] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0025] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0026] In the technical solution of this invention, the collection and use of data are carried out with the user's permission and in accordance with relevant regulations.

[0027] Example 1, as Figure 1 As shown, this embodiment of the invention provides a method for generating transaction reports for anti-money laundering monitoring, which specifically includes the following steps:

[0028] S10: Monitor the transactions of the target account, calculate and obtain the short-term transaction amount and short-term transaction frequency when there are fluctuations in transaction amount, and perform amount anomaly analysis and frequency anomaly analysis to obtain the target amount anomaly coefficient and the target frequency anomaly coefficient.

[0029] In this embodiment, the target account is the account used for anti-money laundering monitoring, and can be any user's account. When monitoring transactions of the target account, if fluctuations in transaction amounts are detected, such as amounts exceeding normal transaction amounts, it indicates potential money laundering activity, triggering further analysis mechanisms to determine whether the transaction is suspicious.

[0030] Specifically, the short-term transaction amount and short-term transaction frequency of the target account are further calculated and obtained, and the amount anomaly analysis and frequency anomaly analysis are performed respectively. Specifically, the degree of anomaly of the short-term transaction amount and short-term transaction frequency relative to the previous transaction amount and transaction frequency of the target account is analyzed to obtain the target amount anomaly coefficient and the target frequency anomaly coefficient.

[0031] Through the above method, the embodiments of the present invention can effectively capture abnormal fluctuations in the transaction amount and frequency of the account itself. Compared with the traditional method that only relies on static amount thresholds (such as simply judging whether the transaction amount exceeds 100,000 yuan), this method is more dynamic, adaptable and accurate, and provides more reliable basic data support for subsequent split transaction detection.

[0032] Step S10 in the method provided in this application embodiment includes:

[0033] Monitor the transactions of the target account to obtain real-time transaction amounts;

[0034] Determine whether the real-time transaction amount is greater than the fluctuation amount threshold. If yes, a fluctuation transaction amount occurs; otherwise, no fluctuation transaction occurs.

[0035] When fluctuations occur in transaction amounts, the total transaction amount and total number of transactions of the target account within a preset time range are calculated to obtain the short-term transaction amount, and the short-term transaction frequency is calculated based on the total number of transactions.

[0036] Based on the short-term transaction amount and short-term transaction frequency, perform amount anomaly analysis and frequency anomaly analysis to obtain the target amount anomaly coefficient and the target frequency anomaly coefficient.

[0037] In this embodiment, when monitoring transactions of a target account, the transaction amount of the account is first obtained in real time, i.e., the specific transaction amount executed by the account at a certain moment, such as a user completing a transfer transaction of 50,000 yuan. Then, this real-time transaction amount is compared with a preset fluctuation threshold. This threshold can be set by historical transaction data or industry standards. For example, the normal transaction amount of an account usually does not exceed 20,000 yuan. If a transaction of this account exceeds 20,000 yuan, it is determined that the transaction has triggered the condition for a fluctuating transaction amount. If the transaction amount is greater than the threshold, a fluctuating transaction amount occurs, the transaction is recorded and further analyzed; otherwise, it is considered a normal transaction, no fluctuating transaction amount occurs, and no special processing is required.

[0038] For example, the volatility threshold can be set based on the account's historical transaction volume. For instance, the average transaction volume of all transactions within a historical period of the target account can be calculated, multiplied by a volatility coefficient, such as 2, to set a normal range for transaction volume fluctuations. The product is then used as the volatility threshold. The volatility threshold can also be customized based on the account's risk profile.

[0039] When fluctuating transaction amounts are detected, in order to comprehensively evaluate the recent trading behavior of the account, the total transaction amount and total number of transactions of the target account within a preset time range in the past are calculated. The preset time range in the past is, for example, 1 hour. All transaction data of the target account within the preset time range in the past (including transfers in, transfers out, and consumption, etc.) are collected, the total transaction amount is calculated, and the total number of transactions is counted, such as 6 times.

[0040] The total transaction amount within a preset time range is used as the short-term transaction amount, and the ratio of the total number of transactions to the length of the preset time range is calculated as the short-term transaction frequency, for example, 6 times / hour.

[0041] Furthermore, after obtaining the short-term transaction amount and short-term transaction frequency, further analysis of transaction amount anomalies and frequency anomalies is performed to obtain the degree of anomaly of the target account's recent transactions relative to the account's previous regular transactions, and to obtain the target transaction amount anomaly coefficient and the target frequency anomaly coefficient.

[0042] The step of "performing amount anomaly analysis and frequency anomaly analysis based on the short-term transaction amount and short-term transaction frequency to obtain target amount anomaly coefficient and target frequency anomaly coefficient" in the method provided in this application embodiment includes:

[0043] Calculate the fluctuation coefficient between the short-term transaction amount and the fluctuation amount threshold, and use it as the target amount anomaly coefficient;

[0044] Calculate the fluctuation coefficient between the short-term trading frequency and the average trading frequency of the target account, and use it as the target frequency anomaly coefficient.

[0045] In this embodiment, short-term transaction amount fluctuation analysis is first performed, and the degree of transaction abnormality is quantified by calculating the fluctuation coefficient, thereby more accurately identifying suspicious transaction behavior.

[0046] Specifically, the volatility coefficient between the short-term transaction amount and the volatility threshold is calculated and used as the target amount anomaly coefficient. For example, the difference between the short-term transaction amount and the volatility threshold is calculated, and then the ratio of this difference to the volatility threshold is calculated; or the ratio of the short-term transaction amount to the volatility threshold is calculated and used as the target amount anomaly coefficient. The larger this target amount anomaly coefficient, the larger the short-term transaction amount and the greater the degree of anomaly.

[0047] Furthermore, an analysis of abnormal fluctuations in short-term trading frequency is conducted. First, the average trading frequency of the target account over a historical period is calculated, for example, by obtaining all trading data of the target account over a historical period. Then, the average trading frequency within each preset time range is calculated as the average trading frequency, for example, 1 time / hour.

[0048] Then, the volatility coefficient between the short-term trading frequency and the average trading frequency is calculated. For example, the difference between the short-term trading frequency and the average trading frequency is calculated, and then the ratio of this difference to the average trading frequency is calculated, or the ratio of the short-term trading frequency to the average trading frequency is calculated, which serves as the target frequency anomaly coefficient. The larger the target frequency anomaly coefficient, the greater the degree of anomaly in the trading frequency of the current target account. For example, the target amount anomaly coefficient and the target frequency anomaly coefficient for the target account are 1.2 and 1.5, respectively.

[0049] The larger the target amount anomaly coefficient and the target frequency anomaly coefficient are, the greater the probability that the target account's recent transaction behavior is abnormal and the greater the probability that money laundering behavior exists.

[0050] Through the above calculations, the embodiments of the present invention can dynamically quantify the fluctuations in the transaction amount and frequency of an account. Compared with the traditional fixed threshold judgment method, it can more comprehensively analyze the account's transaction patterns, improve the ability to identify abnormal transactions, and provide more accurate data support for subsequent anti-money laundering monitoring.

[0051] S20: Obtain the associated account network of the target account, monitor and calculate the associated short-term transaction amount and associated short-term transaction frequency of multiple associated accounts within the associated account network, and obtain the associated amount network and associated frequency network;

[0052] In this embodiment of the application, since splitting transactions involves breaking down large transactions into multiple smaller transactions to hide the source and increase the difficulty of monitoring, it is necessary to monitor and analyze other related accounts that have transaction behavior with the target account in order to improve the accuracy of monitoring splitting transaction behavior.

[0053] Specifically, the system identifies accounts with financial transactions with the target account, constructs a network of related accounts, and monitors and calculates the short-term transaction amounts and frequencies of multiple related accounts within the network to obtain a network of related amounts and a network of related frequencies.

[0054] In this way, short-term transaction data of all related accounts are collected to construct a network of related amounts and a network of related frequencies, so as to identify hidden splitting transaction behavior based on the transaction data of related accounts and improve the ability to detect hidden money laundering activities.

[0055] Step S20 in the method provided in this application embodiment includes:

[0056] Obtain multiple associated accounts that are related to the target account to obtain a network of associated accounts;

[0057] The monitoring and calculation process obtains the associated short-term transaction amounts and associated short-term transaction frequencies of multiple associated accounts within the associated account network, thus obtaining the associated amount network and the associated frequency network.

[0058] In this embodiment of the application, multiple associated accounts with financial transactions with the target account are obtained through the target account's historical transaction records. Associated accounts refer to accounts that have had frequent transactions with the target account or have a financial transfer link. For example, if target account A transfers funds to accounts B, C, and D, then accounts B, C, and D can be regarded as associated accounts of A, and together they constitute an associated account network.

[0059] Furthermore, based on the steps described above, short-term transaction data within the most recent preset time range of multiple related accounts within the related account network are monitored and obtained, and the related short-term transaction amounts and frequencies of the multiple related accounts are statistically calculated. In this way, a related amount network and a related frequency network of multiple related accounts are obtained.

[0060] Compared to traditional single-account monitoring methods, the embodiments of this application construct a network of related accounts and analyze the related amounts and frequency characteristics, which can serve as a data basis for identifying suspicious behaviors that use multiple accounts to split and conceal transaction patterns.

[0061] S30: Based on the short-term transaction amount, predict the split transaction compensation ratio and the number of splits to obtain the split transaction compensation ratio and the number of splits;

[0062] In this embodiment of the application, after calculating the short-term transaction amount, the prediction of the split transaction compensation ratio and the number of splits is further introduced to identify potential split transaction behavior and improve the accuracy of anti-money laundering monitoring.

[0063] Splitting transactions involves dividing large sums of money into multiple smaller transactions and adding funds from other sources to these smaller transactions. This bypasses traditional anti-money laundering monitoring rules and increases the difficulty of detection. For example, if an account needs to transfer 200,000 yuan, a direct transfer might trigger large transaction monitoring. Therefore, it might be split into 10 transactions, with 2,000 yuan added to each, forming a transaction of 22,000 yuan per transaction to circumvent regulations. Thus, monitoring split transactions involves adjusting the total amount and increasing the number of transactions. This application's embodiments are based on this approach to monitor suspicious split transaction behavior and generate transaction reports.

[0064] Among them, the compensation adjustment and the number of splits for split transactions of different amounts have certain data patterns. For example, the larger the amount of the split transaction, the greater the compensation adjustment and the more splits, in order to avoid monitoring.

[0065] Therefore, this application embodiment predicts the compensation ratio and number of split transactions based on the short-term transaction amount, and obtains the compensation ratio and number of split transactions for the current target account to carry out split transactions, so as to serve as the data basis for subsequent monitoring of related account transactions, and to carry out more accurate and evidence-based split transaction monitoring.

[0066] This invention, by predicting the compensation ratio and the number of split transactions based on short-term transaction amounts, can effectively identify more covert splitting patterns and improve the intelligence level of anti-money laundering monitoring.

[0067] Step S30 in the method provided in this application embodiment includes:

[0068] Based on anti-money laundering monitoring data over a historical period, a set of sample short-term transaction amounts was collected, and the split transaction compensation ratio and split quantity of each sample short-term transaction amount in the split transaction were obtained and labeled as the sample split transaction compensation ratio set and sample split quantity set.

[0069] The sample short-term transaction amount set is used as the input feature, and the sample split transaction compensation ratio set and sample split quantity set are used as the output features to train a split transaction feature predictor.

[0070] The short-term transaction amount is input into the split transaction feature predictor to obtain the split transaction compensation ratio and the number of splits.

[0071] In this embodiment of the application, a split transaction feature prediction model is constructed based on anti-money laundering monitoring data over a historical period to accurately identify split transaction behavior.

[0072] Specifically, based on transaction data monitored and ultimately determined to be money laundering split transactions within a historical period, a sample set of short-term transaction amounts is collected. This involves selecting short-term transaction amounts from accounts involved in money laundering split transactions from historical transaction data as sample short-term transaction amounts. For example, sample short-term transaction amounts might be 50,000 yuan, 120,000 yuan, 300,000 yuan, etc.

[0073] Furthermore, the compensation ratio and number of splits are obtained for the corresponding money laundering split transactions that split the sample short-term transaction amount. The compensation ratio refers to the increase in the actual total transaction amount after splitting compared to the original transaction amount, and the number of splits is the number of small transactions that split the original sample short-term transaction amount. Both are characteristic parameters of the split transaction.

[0074] For example, if an account has a short-term transaction amount of 50,000 yuan, and this amount is split and transferred to 10 related accounts, the total transfer amount is 55,000 yuan. The compensation ratio for this split transaction is 5,000 yuan / 50,000 yuan = 10%, and the number of split transactions is 10. Thus, a set of compensation ratios and a set of the number of split transactions are obtained. For example, the short-term transaction amount, the compensation ratio, and the number of split transactions might include: short-term transaction amount: 50,000 yuan, number of split transactions: 4, compensation ratio: 1.1; and short-term transaction amount: 200,000 yuan, number of split transactions: 8, compensation ratio: 1.2.

[0075] By collecting a set of sample split transaction compensation ratios and a set of sample split transaction quantities from historical money laundering split transactions, we can provide data support for subsequent training to predict split transaction compensation ratios and split transaction quantities.

[0076] Furthermore, a set of sample short-term transaction amounts is used as input features, and a set of sample split transaction compensation ratios and a set of sample split quantities are used as output features to train a split transaction feature predictor. This split transaction feature predictor can be trained using a feedforward neural network in machine learning, predicting split transaction features by learning from historical split transaction data.

[0077] For example, a feedforward neural network is used to train a split transaction feature predictor to accurately identify the split transaction compensation ratio and the number of splits. The feedforward neural network constructs the split transaction feature predictor, which includes an input layer, hidden layers, and an output layer. The input layer has a dimension of 1, representing the short-term transaction amount, while the output layer has a dimension of 2, corresponding to the split transaction compensation ratio and the number of splits. A linear activation function is used. During training, the sample short-term transaction amount is input, and the output split transaction compensation ratio and the number of splits are obtained. The difference between these values ​​and the corresponding real data (i.e., the sample split transaction compensation ratio and the sample split number) is calculated, and the loss is calculated using a mean squared error loss function. Then, the network parameters of the hidden layer, such as the weights, are tuned to reduce the loss, making the predicted value closer to the actual split feature data of historical data, i.e., reducing the loss. This iterative training is performed on all sample data until requirements are met, such as a loss less than a preset value (e.g., 5%) and an accuracy greater than 95%, at which point training is complete, and the split transaction feature predictor is obtained.

[0078] Inputting the short-term transaction amount of the current target account into the split transaction feature predictor yields the predicted split transaction compensation ratio and split quantity. This reflects the possible compensation ratio and split quantity of the target account for splitting short-term transaction amounts, which is used to monitor and determine subsequent money laundering activities through split transactions, thereby improving the accuracy and reliability of monitoring.

[0079] S40: Based on the associated amount network and associated frequency network, perform associated transaction anomaly analysis using the split transaction compensation ratio and split quantity to obtain associated amount anomaly coefficient and associated frequency anomaly coefficient. Combine the target amount anomaly coefficient and target frequency anomaly coefficient to calculate the total anomaly coefficient and generate a transaction report.

[0080] In this embodiment, based on the associated amount network and the associated frequency network, combined with the split transaction compensation ratio and the number of splits, anomaly analysis of associated transactions is performed, the associated amount anomaly coefficient and the associated frequency anomaly coefficient are calculated, and finally the total anomaly coefficient is calculated by combining the target amount anomaly coefficient and the target frequency anomaly coefficient, and a transaction report is generated to improve the accuracy and comprehensiveness of anti-money laundering monitoring.

[0081] The associated amount network and associated frequency network include the actual amount and frequency of recent transactions of multiple associated accounts. Based on the short-term transaction amount and frequency of the target account, combined with the split transaction compensation ratio and the number of splits, the amount and frequency of split transactions can be predicted. By combining the associated amount network and associated frequency network for verification, the suspicious probability of multiple associated accounts conducting split transactions can be obtained, and then the associated amount anomaly coefficient and associated frequency anomaly coefficient can be calculated.

[0082] Then, by combining the target amount anomaly coefficient and the target frequency anomaly coefficient of the target account itself, the total anomaly coefficient is calculated, and a transaction report is generated.

[0083] Step S40 in the method provided in this application embodiment includes:

[0084] Calculate the sum of multiple related short-term transaction amounts within the related amount network to obtain the actual split amount;

[0085] The compensation for the short-term transaction amount is calculated using the aforementioned split transaction compensation ratio to obtain the predicted split amount;

[0086] Calculate the sum of multiple associated short-term transaction frequencies within the associated frequency network to obtain the actual split transaction frequency;

[0087] The predicted split trading frequency is obtained by multiplying the number of splits by the short-term trading frequency.

[0088] Calculate the similarity between the actual split amount and the predicted split amount to obtain the correlation amount anomaly coefficient; calculate the similarity between the actual split transaction frequency and the predicted split transaction frequency to obtain the correlation frequency anomaly coefficient.

[0089] In this embodiment of the application, the sum of multiple related short-term transaction amounts within the related amount network is calculated to obtain the actual split amount, which is the total amount of transactions conducted by multiple related accounts within the most recent preset time range.

[0090] Furthermore, the compensation for short-term transactions in the target account is calculated using the split transaction compensation ratio to obtain the predicted split amount after splitting the short-term transaction amount and adding funds from other sources if the target account engages in split transactions. For example, the predicted split amount is obtained by adding 1 to the split transaction compensation ratio and multiplying it by the short-term transaction amount.

[0091] If the target account and multiple related accounts engage in money laundering through split transactions, then the related accounts should have recently only transacted with the target account. In this case, the actual split amount should be close to the predicted split amount. If the target account and multiple related accounts do not engage in split transactions, then the related accounts are simply conducting normal transactions. In this case, the actual split amount is likely to differ significantly from the predicted split amount. Therefore, calculating the similarity between the actual split amount and the predicted split amount can yield the probability of suspicious split transactions between the target account and multiple related accounts, serving as an anomaly coefficient for related amount transactions.

[0092] For example, the ratio of the absolute value of the difference between the actual split amount and the predicted split amount to the actual split amount is calculated. Then, 1 is subtracted from this ratio to obtain the similarity score. The smaller the difference, the greater the similarity score, and the higher the probability of a split transaction. This similarity score is used as an anomaly coefficient for related amounts.

[0093] Furthermore, the sum of multiple short-term transaction frequencies of multiple associated accounts within the associated frequency network is calculated to obtain the actual split transaction frequency. The actual split transaction frequency is the total frequency of transactions conducted by multiple associated accounts within a recent preset time range.

[0094] Furthermore, the predicted split trading frequency is obtained by multiplying the number of splits by the short-term trading frequency. The predicted split trading frequency is the total trading frequency of the multiple related accounts if the target account and multiple related accounts conduct split trading. If each transaction of the target account is split according to the number of splits, then the total trading frequency of split trading is the short-term trading frequency multiplied by the number of splits.

[0095] If the target account and multiple associated accounts engage in split trading, then these associated accounts should only be trading with the target account recently. Therefore, the sum of the trading frequencies of these associated accounts should be the ratio of the split transaction amount (after splitting recent transaction amounts) to the time frame. In other words, the actual split trading frequency and the predicted split trading frequency should be relatively close. If the target account and multiple associated accounts do not engage in split trading, then the trading frequencies of the associated accounts and the target account are completely random, and the actual split trading frequency and the predicted split trading frequency should differ significantly.

[0096] Therefore, the similarity between the actual split transaction frequency and the predicted split transaction frequency is calculated to obtain the correlation frequency anomaly coefficient indicating that multiple related accounts and target accounts are engaging in split transactions for money laundering. For example, the ratio of the absolute value of the difference between the actual and predicted split transaction frequencies to the predicted split transaction frequency is calculated, and 1 minus this ratio is used as the similarity, i.e., the correlation frequency anomaly coefficient. The smaller the difference, the greater the similarity, the greater the probability that multiple related accounts and target accounts are engaging in split transactions, and the larger the correlation frequency anomaly coefficient.

[0097] This invention, through analysis of associated account networks, prediction of split transaction characteristics and transaction data, and calculation of anomaly coefficients, can dynamically identify complex split transaction money laundering patterns, thereby improving the intelligence and accuracy of monitoring.

[0098] In this embodiment of the application, based on the calculated correlation amount anomaly coefficient and correlation frequency anomaly coefficient, combined with the target amount anomaly coefficient and target frequency anomaly coefficient of the target account, the total anomaly coefficient of the current target account that has split transactions is analyzed and calculated to reflect the overall anomaly probability and generate a monitoring transaction report.

[0099] Step S40 in the method provided in this application embodiment further includes:

[0100] The total abnormal coefficient is calculated based on the abnormal coefficients of the associated amount, the abnormal coefficient of the associated frequency, the abnormal coefficient of the target amount, and the abnormal coefficient of the target frequency.

[0101] A transaction report is generated based on the short-term transaction amount, short-term transaction frequency, and associated account network, combined with the total anomaly coefficient.

[0102] In this embodiment, a total anomaly coefficient is calculated based on the associated amount anomaly coefficient, the associated frequency anomaly coefficient, the target amount anomaly coefficient, and the target frequency anomaly coefficient. For example, the average of the associated amount anomaly coefficient, the associated frequency anomaly coefficient, the target amount anomaly coefficient, and the target frequency anomaly coefficient is calculated to obtain the total anomaly coefficient. The larger the total anomaly coefficient, the greater the probability and scale of money laundering through split transactions in the current target account.

[0103] Furthermore, based on the monitored short-term transaction amounts, short-term transaction frequencies, and related account networks, combined with the total anomaly coefficient, a transaction report for monitoring money laundering through split transactions is generated, which is more comprehensive and accurate.

[0104] The transaction report generation method for anti-money laundering monitoring provided in this invention has at least the following technical effects:

[0105] This invention improves the ability to identify split transactions by conducting in-depth analysis of the target account's transaction behavior, thereby effectively reducing the inaccuracy, incompleteness, and false alarms in traditional transaction monitoring reports. First, when monitoring the target account's transactions, short-term transaction amounts and frequencies are calculated, and anomaly analysis of amounts and frequencies is performed to obtain target amount anomaly coefficients and target frequency anomaly coefficients. This allows for accurate identification of short-term transaction fluctuations, offering greater flexibility and adaptability compared to traditional methods that rely solely on fixed amount thresholds. Second, by acquiring the target account's associated account network and monitoring and calculating the associated short-term transaction amounts and frequencies of multiple associated accounts, an associated amount network and associated frequency network are formed. This technical feature effectively identifies multiple associated accounts involved in split transactions, avoiding monitoring blind spots caused by analyzing only a single account, and is applicable to the characteristics of split transactions. Furthermore, by predicting the split transaction compensation ratio and split quantity based on short-term transaction amounts, changes in funds and transaction frequency caused by split transactions can be effectively predicted, making the monitoring and analysis more accurate. Subsequently, anomaly analysis of related transactions is conducted using related amount networks and related frequency networks, combined with the compensation ratio and number of split transactions. This yields anomaly coefficients for related amount and frequency, quantifying the suspiciousness of split transactions and improving the accuracy of anomaly identification. Finally, the total anomaly coefficient is calculated by combining the target amount and target frequency anomaly coefficients. A transaction report is then generated based on short-term transaction amounts, short-term transaction frequencies, and related account networks, ensuring more comprehensive report content and enhancing the ability to accurately monitor money laundering activities related to split transactions, thereby improving the accuracy and effectiveness of anti-money laundering monitoring.

[0106] Example 2, as Figure 2 As shown, following the same inventive concept as the transaction report generation method for anti-money laundering monitoring in Embodiment 1, this embodiment of the invention also provides a transaction report generation system for anti-money laundering monitoring. The explanation of the transaction report generation method for anti-money laundering monitoring in Embodiment 1 also applies to the transaction report generation system for anti-money laundering monitoring. The system includes:

[0107] The target anomaly analysis module 11 is used to monitor the transactions of the target account. When there are fluctuations in the transaction amount, it calculates and obtains the short-term transaction amount and short-term transaction frequency, and performs anomaly analysis of the amount and frequency to obtain the target amount anomaly coefficient and the target frequency anomaly coefficient.

[0108] The associated network acquisition module 12 is used to acquire the associated account network of the target account, monitor and calculate the associated short-term transaction amount and associated short-term transaction frequency of multiple associated accounts in the associated account network, and obtain the associated amount network and the associated frequency network.

[0109] The splitting feature prediction module 13 is used to predict the splitting transaction compensation ratio and the number of splitting transactions based on the short-term transaction amount, and to obtain the splitting transaction compensation ratio and the number of splitting transactions.

[0110] The report generation module 14 is used to perform anomaly analysis of related transactions based on the related amount network and the related frequency network, using the split transaction compensation ratio and the number of split transactions, to obtain the related amount anomaly coefficient and the related frequency anomaly coefficient, and to calculate the total anomaly coefficient by combining the target amount anomaly coefficient and the target frequency anomaly coefficient, and to generate a transaction report.

[0111] Furthermore, the target anomaly analysis module 11 is also used for:

[0112] Monitor the transactions of the target account to obtain real-time transaction amounts;

[0113] Determine whether the real-time transaction amount is greater than the fluctuation amount threshold. If yes, a fluctuation transaction amount occurs; otherwise, no fluctuation transaction occurs.

[0114] When fluctuations occur in transaction amounts, the total transaction amount and total number of transactions of the target account within a preset time range are calculated to obtain the short-term transaction amount, and the short-term transaction frequency is calculated based on the total number of transactions.

[0115] Based on the short-term transaction amount and short-term transaction frequency, perform amount anomaly analysis and frequency anomaly analysis to obtain the target amount anomaly coefficient and the target frequency anomaly coefficient.

[0116] Specifically, based on the short-term transaction amount and short-term transaction frequency, anomaly analysis of the transaction amount and frequency are performed to obtain target transaction amount anomaly coefficients and target frequency anomaly coefficients, including:

[0117] Calculate the fluctuation coefficient between the short-term transaction amount and the fluctuation amount threshold, and use it as the target amount anomaly coefficient;

[0118] Calculate the fluctuation coefficient between the short-term trading frequency and the average trading frequency of the target account, and use it as the target frequency anomaly coefficient.

[0119] Furthermore, the associated network acquisition module 12 is also used for:

[0120] Obtain multiple associated accounts that are related to the target account to obtain a network of associated accounts;

[0121] The monitoring and calculation process obtains the associated short-term transaction amounts and associated short-term transaction frequencies of multiple associated accounts within the associated account network, thus obtaining the associated amount network and the associated frequency network.

[0122] Furthermore, the split feature prediction module 13 is also used for:

[0123] Based on anti-money laundering monitoring data over a historical period, a set of sample short-term transaction amounts was collected, and the split transaction compensation ratio and split quantity of each sample short-term transaction amount in the split transaction were obtained and labeled as the sample split transaction compensation ratio set and sample split quantity set.

[0124] The sample short-term transaction amount set is used as the input feature, and the sample split transaction compensation ratio set and sample split quantity set are used as the output features to train a split transaction feature predictor.

[0125] The short-term transaction amount is input into the split transaction feature predictor to obtain the split transaction compensation ratio and the number of splits.

[0126] Furthermore, the report generation module 14 is also used for:

[0127] Calculate the sum of multiple related short-term transaction amounts within the related amount network to obtain the actual split amount;

[0128] The compensation for the short-term transaction amount is calculated using the aforementioned split transaction compensation ratio to obtain the predicted split amount;

[0129] Calculate the sum of multiple associated short-term transaction frequencies within the associated frequency network to obtain the actual split transaction frequency;

[0130] The predicted split trading frequency is obtained by multiplying the number of splits by the short-term trading frequency.

[0131] Calculate the similarity between the actual split amount and the predicted split amount to obtain the correlation amount anomaly coefficient; calculate the similarity between the actual split transaction frequency and the predicted split transaction frequency to obtain the correlation frequency anomaly coefficient.

[0132] Monitor the transaction data of the target account and obtain the transaction data sequence within the most recent preset time range;

[0133] Extract the transaction amount and transaction timestamp of multiple transaction data within the transaction data sequence to obtain the transaction amount sequence and the transaction timestamp sequence;

[0134] The sum of the transaction amount sequence is calculated to obtain the short-term transaction amount, and the short-term transaction frequency is calculated based on the transaction timestamp sequence.

[0135] Furthermore, the report generation module 14 is also used for:

[0136] The total abnormal coefficient is calculated based on the abnormal coefficients of the associated amount, the abnormal coefficient of the associated frequency, the abnormal coefficient of the target amount, and the abnormal coefficient of the target frequency.

[0137] A transaction report is generated based on the short-term transaction amount, short-term transaction frequency, and associated account network, combined with the total anomaly coefficient.

[0138] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0139] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention 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.

[0140] This invention is described with reference to flow diagrams and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and / or block in the flow diagram and / or block diagram, and combinations of flow diagrams 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, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate flow diagrams and / or block diagrams for implementing flow diagrams and / or block diagrams. Figure 1 One or more traffic and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0141] 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 implements the flow Figure 1 One or more traffic and / or boxes Figure 1 The function specified in one or more boxes.

[0142] 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 in the flow. Figure 1 One or more traffic and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0143] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0144] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for generating transaction reports for anti-money laundering monitoring, characterized in that, The method includes: The transaction monitoring of the target account is carried out. When the transaction amount fluctuates, the short-term transaction amount and short-term transaction frequency are calculated and obtained. The amount anomaly analysis and frequency anomaly analysis are performed to obtain the target amount anomaly coefficient and the target frequency anomaly coefficient. Obtain the network of associated accounts of the target account, monitor and calculate the associated short-term transaction amount and associated short-term transaction frequency of multiple associated accounts within the network of associated accounts, and obtain the associated amount network and associated frequency network; Based on the short-term transaction amount, the compensation ratio and number of split transactions are predicted to obtain the compensation ratio and number of split transactions. Based on the aforementioned related amount network and related frequency network, anomaly analysis of related transactions is performed using the split transaction compensation ratio and split quantity to obtain the related amount anomaly coefficient and related frequency anomaly coefficient. Combining the target amount anomaly coefficient and target frequency anomaly coefficient, a total anomaly coefficient is calculated, and a transaction report is generated, including: Calculate the sum of multiple related short-term transaction amounts within the related amount network to obtain the actual split amount; The compensation for the short-term transaction amount is calculated using the aforementioned split transaction compensation ratio to obtain the predicted split amount; Calculate the sum of multiple associated short-term transaction frequencies within the associated frequency network to obtain the actual split transaction frequency; The predicted split trading frequency is obtained by multiplying the number of splits by the short-term trading frequency. Calculate the similarity between the actual split amount and the predicted split amount to obtain the correlation amount anomaly coefficient; calculate the similarity between the actual split transaction frequency and the predicted split transaction frequency to obtain the correlation frequency anomaly coefficient.

2. The transaction report generation method for anti-money laundering monitoring according to claim 1, characterized in that, Transaction monitoring is performed on the target account. When fluctuations in transaction amounts occur, short-term transaction amounts and short-term transaction frequencies are calculated and analyzed for anomalies in both amount and frequency. This yields target amount anomaly coefficients and target frequency anomaly coefficients, including: Monitor the transactions of the target account to obtain real-time transaction amounts; Determine whether the real-time transaction amount is greater than the fluctuation amount threshold. If yes, a fluctuation transaction amount occurs; otherwise, no fluctuation transaction occurs. When fluctuations occur in transaction amounts, the total transaction amount and total number of transactions of the target account within a preset time range are calculated to obtain the short-term transaction amount, and the short-term transaction frequency is calculated based on the total number of transactions. Based on the short-term transaction amount and short-term transaction frequency, perform amount anomaly analysis and frequency anomaly analysis to obtain the target amount anomaly coefficient and the target frequency anomaly coefficient.

3. The transaction report generation method for anti-money laundering monitoring according to claim 2, characterized in that, Based on the aforementioned short-term transaction amounts and frequencies, anomaly analysis of transaction amounts and frequencies is performed to obtain target transaction amount anomaly coefficients and target frequency anomaly coefficients, including: Calculate the fluctuation coefficient between the short-term transaction amount and the fluctuation amount threshold, and use it as the target amount anomaly coefficient; Calculate the fluctuation coefficient between the short-term trading frequency and the average trading frequency of the target account, and use it as the target frequency anomaly coefficient.

4. The method for generating transaction reports for anti-money laundering monitoring according to claim 1, characterized in that, Obtain the network of associated accounts for the target account, monitor and calculate the associated short-term transaction amounts and frequencies of multiple associated accounts within the network, and obtain the associated amount network and the associated frequency network, including: Obtain multiple associated accounts that are related to the target account to obtain a network of associated accounts; The monitoring and calculation process obtains the associated short-term transaction amounts and associated short-term transaction frequencies of multiple associated accounts within the associated account network, thus obtaining the associated amount network and the associated frequency network.

5. The transaction report generation method for anti-money laundering monitoring according to claim 1, characterized in that, Based on the short-term transaction amount, the compensation ratio and number of split transactions are predicted to obtain the compensation ratio and number of split transactions, including: Based on anti-money laundering monitoring data over a historical period, a set of sample short-term transaction amounts was collected, and the split transaction compensation ratio and split quantity of each sample short-term transaction amount in the split transaction were obtained and labeled as the sample split transaction compensation ratio set and sample split quantity set. The sample short-term transaction amount set is used as the input feature, and the sample split transaction compensation ratio set and sample split quantity set are used as the output features to train a split transaction feature predictor. The short-term transaction amount is input into the split transaction feature predictor to obtain the split transaction compensation ratio and the number of splits.

6. The method for generating transaction reports for anti-money laundering monitoring according to claim 1, characterized in that, By combining the target amount anomaly coefficient and the target frequency anomaly coefficient, a total anomaly coefficient is calculated, and a transaction report is generated, including: The total abnormal coefficient is calculated based on the abnormal coefficients of the associated amount, the abnormal coefficient of the associated frequency, the abnormal coefficient of the target amount, and the abnormal coefficient of the target frequency. A transaction report is generated based on the short-term transaction amount, short-term transaction frequency, and associated account network, combined with the total anomaly coefficient.

7. A transaction report generation system for anti-money laundering monitoring, characterized in that, The system includes: The target anomaly analysis module is used to monitor the transactions of target accounts. When fluctuations occur in transaction amounts, it calculates and obtains short-term transaction amounts and short-term transaction frequencies, and performs amount anomaly analysis and frequency anomaly analysis to obtain the target amount anomaly coefficient and the target frequency anomaly coefficient. The associated network acquisition module is used to acquire the associated account network of the target account, monitor and calculate the associated short-term transaction amount and associated short-term transaction frequency of multiple associated accounts within the associated account network, and obtain the associated amount network and associated frequency network. The splitting feature prediction module is used to predict the splitting transaction compensation ratio and the number of splits based on the short-term transaction amount, and to obtain the splitting transaction compensation ratio and the number of splits. The report generation module is used to perform anomaly analysis of related transactions based on the related amount network and related frequency network, using the split transaction compensation ratio and split quantity, to obtain the related amount anomaly coefficient and related frequency anomaly coefficient. Combining the target amount anomaly coefficient and target frequency anomaly coefficient, it calculates the total anomaly coefficient and generates a transaction report, including: Calculate the sum of multiple related short-term transaction amounts within the related amount network to obtain the actual split amount; The compensation for the short-term transaction amount is calculated using the aforementioned split transaction compensation ratio to obtain the predicted split amount; Calculate the sum of multiple associated short-term transaction frequencies within the associated frequency network to obtain the actual split transaction frequency; The predicted split trading frequency is obtained by multiplying the number of splits by the short-term trading frequency. Calculate the similarity between the actual split amount and the predicted split amount to obtain the correlation amount anomaly coefficient; calculate the similarity between the actual split transaction frequency and the predicted split transaction frequency to obtain the correlation frequency anomaly coefficient.

Images