Financial data generation method for illegal fund-raising transaction risk scenarios

Abstract

The application provides a financial data generation method for illegal fund-raising transaction risk scenarios, comprising the steps of generating user basic data, generating a user relationship network, and generating normal transaction data between users and multi-scenario illegal fund-raising transaction data. Since user information is generated according to statistical data, and a relationship network between multiple users is constructed on this basis, the real social and economic characteristics of multiple users can be well simulated, diversified and realistic user basic data and a close-to-real social relationship network are provided, and a solid foundation is provided for the simulation of transaction data. Furthermore, based on user basic information, a user relationship network, statistical data and predetermined generation rules, normal transaction data and illegal fund-raising transaction data for multiple illegal fund-raising transaction scenarios are generated, that is, transaction data with complex transaction modes can be automatically generated, thereby providing strong support for financial model testing and fraud detection applications.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of table data generation, and particularly relates to a financial data generation method for illegal fund-raising transaction risk scenarios. BACKGROUND

[0002] With the increasing strictness of data privacy protection regulations, it is extremely difficult to obtain and use real user financial transaction data. Especially in the financial field, user transaction data often involves sensitive personal privacy and is protected by strict legal restrictions and privacy policies, so financial institutions cannot provide or disclose these data. In this context, financial institutions and researchers generally face difficulties in obtaining data when conducting data analysis, model training and fraud detection in the financial field, which greatly reduces the efficiency of risk monitoring and fraud detection. Although there are currently some privacy computing technologies, such as homomorphic encryption, federated learning and differential privacy, these technologies can ensure privacy protection to some extent during data exchange and analysis, and ensure that data is calculated without exposure, but the core problem of not being able to obtain original data has not been solved. In many cases, even if these technologies can guarantee the security and privacy of data during exchange, financial institutions and other related parties are still reluctant to disclose or share real transaction data due to concerns about privacy leaks, data misuse or compliance issues. Therefore, when conducting large-scale and complex financial data analysis, data acquisition still faces great challenges.

[0003] At present, although there are data synthesis methods, mainly based on generative models to synthesize synthetic data close to real data in features, such as based on generative adversarial networks, variational autoencoders, diffusion models, etc. These methods have achieved certain results in specific application scenarios, but in these methods, the generative model used still relies on a large amount of real original data for training. In addition, the model often has difficulty in completely capturing the complexity of the actual scenario when generating data, especially in the financial transaction field, real data contains complex behavior patterns, multi-level network relationships and nonlinear correlations, and these details are often not well represented in the data synthesized by existing data synthesis methods. In the research of illegal fund-raising risk detection, due to the above-mentioned problems of existing data synthesis methods, the synthetic data is difficult to reflect multiple illegal fund-raising transaction patterns, which affects the efficiency of the research and the accuracy of the corresponding detection method designed.

[0004] Therefore, in view of the above problems and shortcomings, there is an urgent need for an illegal fund-raising transaction data synthesis method that does not rely on real data to generate high-quality transaction data, thereby assisting in the modeling research of illegal fund-raising risk detection. SUMMARY

[0005] The present application is carried out to solve the above problems, and aims to provide a financial data synthesis method which is independent of real data and can synthesize financial data containing complex behavior patterns, multi-level network relationships and nonlinear correlations.

[0006] The present application provides a financial data generation method for illegal fund-raising transaction risk scenarios, which has the following technical features: step S1, generating user basic data of a plurality of users; step S2, constructing a user relationship network of the plurality of users based on the user basic data; step S3, generating normal transaction data between the users based on the user basic information, the user relationship network, statistical data and predetermined normal transaction generation rules; and step S4, generating illegal fund-raising transaction data of a plurality of illegal fund-raising transaction scenarios between the users based on the user basic information, the user relationship network, statistical data and predetermined illegal fund-raising transaction generation rules, wherein the user basic data of each user includes bank card information and user information, the user information at least contains age, income level and occupation group, and is generated according to corresponding statistical data respectively, the user relationship network contains a plurality of nodes and edges between the nodes, the nodes are the users, and the edges are relationships between the users.

[0007] The financial data generation method for illegal fund-raising transaction risk scenarios provided by the present application can also have the following technical features: the bank card information includes bank card number, bank affiliation and balance, the user information further includes name, registration date, gender and suspicious object marker, in step S1, for the bank card information, the bank card number is generated by a unique identifier, and the balance is randomly set within a predetermined value range according to the corresponding income level, and for the user information, the income level and the occupation group are distributed according to the proportion of age distribution based on corresponding statistical data.

[0008] The financial data generation method for illegal fund-raising transaction risk scenarios provided by the present application can also have the following technical features: the user relationship network contains a plurality of nodes and edges between the nodes, the nodes are the users, and the edges are relationships between the users, the types of the relationships include parent-child relationship, relative relationship and friend relationship, the friend relationship includes general friend relationship and better friend relationship, and in step S2, an initial user relationship network is generated based on the user basic data and the types of the relationships and is dynamically updated.

[0009] The financial data generation method for illegal fund-raising transaction risk scenarios provided by the application can also have the following technical features: in step S2, for the friend relationship, the potential friend group of the target user is determined according to the age difference between the target user and the users; two groups of users of a predetermined number are randomly selected from the potential friend group, and are respectively set as friends with general relationship and friends with better relationship of the target user, and the bidirectionality of the friend relationship is ensured; a predetermined number of users are selected from the friends of the friends of the target user, and are also set as friends of the target user; for the relative relationship, a predetermined number of users are randomly selected for the target user, and are set as relatives of the target user, and the bidirectionality of the relative relationship is ensured; for the parent relationship, the users are selected for the target user according to the age difference between the target user and the users and a predetermined logical rule, and are set as father or mother, and the uniqueness of the parent relationship is ensured.

[0010] The financial data generation method for illegal fund-raising transaction risk scenarios provided by the application can also have the following technical features: in step S3, the user basic data of a plurality of users is statistically analyzed to obtain feature data; the user group to which the plurality of users belong is determined according to the feature data; the normal transaction generation rule of the user group is generated according to the determined user group and a plurality of predetermined transaction data characteristic information corresponding to the user group; and the normal transaction data between the plurality of users is generated based on the user basic data and the normal transaction generation rule.

[0011] The financial data generation method for illegal fund-raising transaction risk scenarios provided by the application can also have the following technical features: the transaction data characteristic information includes transaction object range, transaction time range, transaction amount range, transaction frequency range, and transaction intention range, and is set according to the age, income level, and occupation group of the users in the user group.

[0012] The financial data generation method for illegal fund-raising transaction risk scenarios provided by the application can also have the following technical features: in step S4, based on the operation mode under the illegal fund-raising transaction scenario, a plurality of users with relationship edges are selected from the user relationship network and an illegal fund-raising relationship network is constructed; the illegal fund-raising transaction generation rule is generated based on the transaction data characteristic information under the illegal fund-raising transaction scenario; and the illegal fund-raising transaction data between the plurality of users is generated based on the user basic data of the users in the illegal fund-raising relationship network and the illegal fund-raising transaction generation rule.

[0013] The financial data generation method for illegal fund-raising transaction risk scenarios provided by the application can also have the following technical features: the illegal fund-raising transaction scenario is a pyramid selling illegal fund-raising transaction scenario; in step S4, a user is randomly selected as a top-level user from the user relationship network according to the operation mode specific to the pyramid selling illegal fund-raising transaction scenario; one or more of the friends, relatives, and parents of the top-level user are selected as first-level subordinate users of the top-level user, and an upper and lower level relationship edge is formed between the top-level user and the first-level subordinate user; one or more of the friends, relatives, and parents of the first-level subordinate user are selected as second-level subordinate users, and an upper and lower level relationship edge is formed between the first-level subordinate user and the corresponding second-level subordinate user; and the above process is repeated to form a multi-level pyramid illegal fund-raising relationship network.

[0014] The financial data generation method for illegal fund-raising transaction risk scenarios provided by the application can also have the following technical features: the transaction data characteristic information includes transaction object range, transaction time range, transaction amount range, and transaction frequency range; according to the operation mode of the pyramid selling illegal fund-raising transaction scenario, the transaction object range includes the superior user and the subordinate user of the user, and the proportion range of the number of payment and collection transactions between the user and the superior user and between the user and the subordinate user is set respectively; and the transaction frequency range is one to several times per week or per month.

[0015] The financial data generation method for illegal fund-raising transaction risk scenarios provided by the application can also have the following technical features: the illegal fund-raising transaction scenario includes a pyramid selling illegal fund-raising transaction scenario, a virtual currency illegal fund-raising transaction scenario, a real estate illegal fund-raising transaction scenario, an old-age project illegal fund-raising transaction scenario, and an internet financial platform illegal fund-raising transaction scenario.

[0016] Effects of the application

[0017] The financial data generation method for illegal fund-raising transaction risk scenarios provided by the application comprises the steps of generating user basic data, generating a user relationship network, and further generating normal transaction data and various modes of illegal fund-raising transaction data between users. Since the user information is generated according to statistical data, the age, income level, and occupation group of the user are included, and the relationship network between multiple users is constructed on this basis, the real social and economic characteristics of multiple users can be well simulated, the user basic data is diversified and has strong reality, and the social relationship network is closer to reality, thereby providing a solid foundation for the simulation of transaction data. Further, the normal transaction data and the illegal fund-raising transaction data of various illegal fund-raising transaction scenarios are generated based on the user basic information, the user relationship network, the statistical data, and the predetermined generation rules, that is, the transaction data with complex transaction modes can be automatically generated, thereby providing strong support for financial model testing and fraud detection and other applications. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 is a flowchart of the financial data generation method for illegal fund-raising transaction risk scenarios in the embodiments of the application;

[0019] Figure 2 is a flowchart of step S2 in the embodiments of the application;

[0020] Figure 3 is a schematic diagram of the user data relationship network constructed in the embodiments of the application;

[0021] Figure 4 is a flowchart of step S3 in the embodiments of the application;

[0022] Figure 5 is a flowchart of step S4 in the embodiments of the application;

[0023] Figure 6 is a schematic diagram of the pyramid selling illegal fund-raising scenario constructed in the embodiments of the application. DETAILED DESCRIPTION

[0024] In order to make the technical means, creative features, purposes and effects achieved by the application easy to understand, the financial data generation method for illegal fund-raising transaction risk scenarios of the application is specifically described below in combination with embodiments and drawings.

[0025] <EMBODIMENT>

[0026] The application provides a financial data generation method for illegal fund-raising transaction risk scenarios, which aims to simulate transaction data of illegal fund-raising behavior, so as to apply the transaction data to financial risk model testing, fraud detection, consumer behavior analysis, and other application aspects.

[0027] Figure 1 is a flowchart of a financial data generation method for the illegal fund-raising transaction risk scenario in this embodiment.

[0028] As shown in Figure 1 , the method comprises the following steps:

[0029] Step S1, generating user basic data of a plurality of users.

[0030] Step S2, constructing a user relationship network of the plurality of users based on the user basic data.

[0031] Step S3, generating normal transaction data between the plurality of users based on the user basic data, the user relationship network, statistical data of each group, and a predetermined normal transaction generation rule.

[0032] Step S4, generating illegal fund-raising transaction data between the plurality of users based on the user basic data, the user relationship network, statistical data of each group, and a predetermined illegal fund-raising transaction generation rule.

[0033] The above steps will be described in detail below.

[0034] Step S1, generating user basic data of a plurality of users.

[0035] Among them, the user basic data includes bank card information and user information (cardholder information).

[0036] The bank card information includes bank card ID (i.e. the unique identification number of the bank card), bank card belonging bank, and balance. Among them, the bank card ID is generated through a unique identifier, ensuring that the bank card ID of all users is unique and not repeated. The belonging bank is randomly set. The balance is randomly set within the corresponding numerical range according to the income level in the cardholder information.

[0037] The user information includes name, age, registration date, gender, income level, occupation group, and suspicious object mark. Among them, the age is randomly set within the range of 18-60 years old, or is allocated and set based on the proportion of each age group in the corresponding statistical data. The income level and the occupation group are allocated based on the corresponding statistical data according to the proportion of age distribution, for example, the occupation group of 18-25 age group includes students, newly employed personnel, freelancers, etc. The suspicious object mark is used to mark whether the user is a suspicious object related to illegal fund-raising transactions. Marked as 0 indicates that the user is a white list user and is not a suspicious object; marked as 1 indicates that the user is a suspicious object.

[0038] The user basic data generated in the above manner can simulate the real social and economic characteristics of multiple users, provide diversified and highly realistic user basic data, and provide a basis for subsequent transaction simulation.

[0039] The following Table 1 shows the relevant fields of the user basic information.

[0040] Table 1 User basic information field table

[0041]

[0042] The following Table 2 shows a data example based on the above field settings.

[0043] Table 2 User basic information data example table

[0044]

[0045]

[0046] Step S2, based on the user basic information, a user relationship network of multiple users is constructed.

[0047] Among them, the relationship type includes friends, relatives and parents, and the friends are further divided into friends with general relationship and friends with better relationship.

[0048] In step S2, the initial user relationship network is generated based on the user basic data and the above relationship type, and is dynamically updated to ensure the authenticity of the relationship between users and the diversity of the network structure.

[0049] Figure 2 is the flow chart of step S2 in this embodiment.

[0050] As Figure 2 shown, in step S2, generating the initial user relationship network specifically includes the following sub-steps:

[0051] Step S2-1, for the friend relationship, according to the age difference between the target user and other multiple users, the potential friend group of the target user is determined. For example, for a target user, multiple other users with an age difference less than a predetermined threshold are considered as the potential friend group of the target user.

[0052] Step S2-2, randomly select two groups of a predetermined number of users from the potential friend group of the target user, and set them as the friends with general relationship and the friends with better relationship of the target user respectively, and ensure the bidirectionality of the friend relationship.

[0053] Step S2-3, for the target user, a predetermined number of users are selected from the friends of the friends, and are also set as the friends of the target user. Thus, the friend relationship in the user relationship network is further expanded, and the connectivity and diversity of the user relationship network are enhanced.

[0054] Step S2-4, for the relative relationship, a predetermined number of other users are randomly selected as the relatives of the target user, and the bidirectionality of the relative relationship is ensured.

[0055] Step S2-5, for the parent relationship, according to the age difference between the target user and other multiple users and a predetermined logical rule, other users are selected as the father and / or mother of the target user, and the uniqueness of the parent relationship is ensured.

[0056] It can be understood that the above operations can be performed on each user respectively, or a part of the users are selected to perform the above operations.

[0057] Figure 3 is a schematic diagram of the user relationship network constructed in this embodiment.

[0058] As shown in Figure 3 , the user relationship network can be represented as a graph structure, in which the nodes are users and the edges are the relationships between users. For a user A, through the above steps, user B is selected as the father / mother of A, and a corresponding relationship edge is formed between A and B; user C is selected as the relative of A and a corresponding relationship edge is formed; and users D and E are selected as the friends of A, and user D is set as a general friend of A and user E is set as a good friend of A, and corresponding relationship edges are formed. It can be understood that the above relationship objects can also be selected for users B-E respectively, thereby forming a complex user relationship network.

[0059] The user relationship network generated in the above manner conforms to the characteristics of a real social network in structure, and provides a solid foundation for subsequent simulation of transaction data of illegal fund-raising scenarios.

[0060] Step S3, based on the user basic information, the user relationship network, the statistical data of each group, and the predetermined normal transaction generation rule, normal transaction data between the multiple users is generated.

[0061] Figure 4 is a flowchart of step S3 in this embodiment.

[0062] As shown in Figure 4 , step S3 specifically includes the following sub-steps:

[0063] Step S3-1, for the selected multiple users, statistical analysis is performed based on the user basic data to obtain feature data.

[0064] For example, statistical analysis is performed on the ages of the multiple users to obtain the age range to which they belong.

[0065] Step S3-2, according to the feature data, the user group to which the multiple users belong is determined from multiple predetermined user groups.

[0066] For example, according to the age range statistically analyzed, the plurality of users are determined as a student group.

[0067] Step S3-3, according to the determined user group and the predetermined plurality of transaction data characteristics information corresponding to the user group, the normal transaction generation rule of the user group is generated.

[0068] For example, for each user group, the transaction object range, transaction time range, transaction amount range, transaction frequency range, transaction intention range, etc. corresponding to the user group are respectively pre-stored, and the transaction generation rule of the user group is generated based on these data ranges. For example, for the student group, the transaction object range is friends, the transaction time range is weekday evening or weekend, the transaction amount range is below 500 yuan, the transaction frequency range is 10-30 times per month, and the transaction intention range includes repayment, mobile payment consumption, etc. Based on these, the corresponding transaction generation rule is generated, for example, a friend of the user is randomly selected as the transaction object, a time point is randomly selected within the selectable time range, a value is randomly selected within the transaction amount range, etc.

[0069] Step S3-4, according to the generated normal transaction generation rule and the user basic data, the normal transaction data between the selected plurality of users is generated.

[0070] For example, the users in the age range of 18-25 years old are taken as the target group for specific description. For this group, the overall data characteristics are as follows:

[0071] Transfer characteristics: Since users in this age group are mainly in the student or early career stage, their income is limited and their expenses are mostly concentrated on daily consumption and social activities. The transfer between friends is mostly used to share the cost of parties, meals or social activities, while the financial transactions with family members are relatively few, and more independent financial processing is involved. Therefore, the characteristics of this user group are frequent small amount transfers, mainly for daily consumption and social activities. Transfers mostly occur among friends and rarely involve family members.

[0072] Transfer time characteristics: Since students and young career groups usually have social activities on weekends or at night, such as dining out, watching movies or participating in group activities. Therefore, the transfer behavior often occurs during these time periods, which coincides with their social time. Therefore, the consumption of this group is mostly concentrated on weekends and evenings, in line with the time of social activities.

[0073] Amount characteristics: Since the income of this age group is generally low, the living expenses are relatively limited, and the amount required for daily expenses such as dining out, online shopping and small loans is usually not very large, basically within a few hundred yuan. Therefore, the consumption of this group is small in single transfer amount, usually within a few hundred yuan.

[0074] Frequency characteristics: Due to frequent social activities and daily consumption, users in this age group have a high number of transfer transactions per month, especially AA-style dining and repayment behaviors make transfer a part of daily life. Therefore, this group has a high consumption frequency, with multiple transfer behaviors per month on average.

[0075] Transfer intent: Due to the fact that young people usually use AA-style payment to share the cost of parties and social activities, in addition, takeout and online shopping are also common consumption methods, and small loans are usually used for emergency financial needs. Therefore, the consumption of this group is mainly for AA-style parties, repayment, mobile payment consumption (such as online shopping and takeout), and small loans. In addition, the professional group also includes students, new employees, and freelancers.

[0076] Taking the student group in the 18-25 age group as an example, the transfer behavior of the student group has significant characteristics, mainly manifested as small and frequent transfers for daily consumption, learning expenses, and social activities. The transfer time is mainly concentrated in the weekend and evening, which is consistent with the work and rest characteristics of students. The single transfer amount is usually small, ranging from 50 yuan to 200 yuan, but the overall frequency is high, with multiple transfer behaviors occurring every week. The main purposes of these transfers include dining, activity expense sharing, online shopping, takeout, and learning-related tuition payments, reflecting the unique consumption needs and social habits of the student group.

[0077] In normal transaction data generation, according to the user's occupation, age group, and income level, etc. Design, combined with the transfer mode, time distribution, amount characteristics, transaction frequency and transaction intent of different professional groups, generate transaction data with high authenticity and representativeness. For example, for the student group, as mentioned above, AA-style consumption is the main feature of their transfer behavior. To simulate this transaction scenario, select a specific time, mainly concentrated in the weekend and evening, select a user as a student, and select some friends of the user to generate consumption behaviors related to dining or online shopping. In this scenario, the student's friends transfer the same amount of money to the student within a few hours, simulating the characteristics of the student group's AA-style consumption in real-life scenarios.

[0078] Step S4, based on user basic information, user relationship network, statistical data of each group, and predetermined illegal fund-raising transaction generation rules, generate illegal fund-raising transaction data between the multiple users.

[0079] Among them, the illegal fund-raising scenarios include pyramid-style illegal fund-raising, virtual currency illegal fund-raising, real estate illegal fund-raising, pension project illegal fund-raising, and Internet financial platform illegal fund-raising. Among them, the most representative is pyramid-style illegal fund-raising.

[0080] Figure 5 is the flowchart of step S4 in this embodiment.

[0081] As shown in Figure 5 Step S4 specifically includes the following sub-steps:

[0082] Step S4-1, based on the operation mode under the illegal fund-raising transaction scenario, selecting multiple users with relationship edges from the user relationship network and constructing an illegal fund-raising relationship network.

[0083] For example, for a pyramid selling illegal fund-raising transaction scenario, according to its unique operation mode, a user is randomly selected from the user relationship network as a top-level user, one or more of the top-level user's friends, relatives, and parents are selected as the top-level user's first-level subordinate users, and the top-level user and the first-level subordinate users form a superior-subordinate relationship edge. Then one or more of the first-level subordinate users' friends, relatives, and parents are selected as the first-level subordinate users' subordinate users, i.e., the top-level user's second-level subordinate users, and the first-level subordinate users and the corresponding second-level subordinate users form a superior-subordinate relationship edge. Repeat the above process to generate a multi-level pyramid illegal fund-raising relationship network.

[0084] Step S4-2, generating illegal fund-raising transaction generation rules based on transaction data characteristic information under the illegal fund-raising transaction scenario.

[0085] For example, for each illegal fund-raising transaction scenario, there are pre-stored transaction object ranges, transaction time ranges, transaction amount ranges, transaction frequency ranges, etc. corresponding to the illegal fund-raising transaction scenario, and based on these data ranges, transaction generation rules between multiple users in the illegal fund-raising relationship network under the scenario are generated.

[0086] For example, for a pyramid selling illegal fund-raising transaction scenario, the transaction object range is the user's superior and subordinate users, and for the superior user, it is mainly to pay, and for the subordinate user, it is mainly to receive (for example, set the proportion range of the number of payments / receipts). The transaction time range is not limited, the transaction amount range is 500-9000 yuan, and the transaction frequency range is once or several times a week, or once or several times a month and embodies periodicity, etc. Based on these predetermined data ranges, illegal fund-raising transaction generation rules are generated.

[0087] Step S4-3, based on the user basic data of the users in the illegal fund-raising relationship network and the illegal fund-raising transaction generation rules, generating illegal fund-raising transaction data between the multiple users.

[0088] For example, take the most representative pyramid selling illegal fund-raising as an example for specific description.

[0089] This mode usually takes fast high returns as bait to achieve illegal fund-raising through multi-level personnel development and fund rolling. For this type of illegal fund-raising, the overall data characteristics are as follows:

[0090] Transfer characteristics: participants in the pyramid scheme will continuously transfer money to their superiors, and the transfer relationship is often complex, involving multiple accounts and intermediaries, and the paths are interlaced.

[0091] Transfer time characteristics: periodic transfers are made on a monthly or weekly basis.

[0092] Amount characteristics: the amount is usually a few hundred to a few thousand yuan, and presents a pyramid structure, with upper-level users mainly as capital inflows, and the inflow amount is large, and lower-level users mainly as capital outflows, and the outflow amount is relatively small.

[0093] Number of times: upper-level users are mainly capital inflows, and the inflow amount (transaction number) is large, while lower-level users are mainly capital outflows, and the outflow is frequent.

[0094] Therefore, based on the above data characteristics of the pyramid illegal fund-raising, first, randomly select a top-level user, and construct a three-level pyramid structure according to the relationships of friends, relatives and parents, and the top-level user absorbs the funds of the lower-level users through the intermediate layer, thereby forming a complete pyramid illegal fund-raising transaction network.

[0095] Figure 6 is a schematic diagram of the pyramid illegal fund-raising scenario constructed in this embodiment.

[0096] As Figure 6 shown, for the pyramid illegal fund-raising scenario in this embodiment, a three-level pyramid structure is constructed according to its unique operation mode, user A is the top-level user, and according to the user relationship network, a number of users (for example, 2-5) are randomly selected from the parents, relatives and friends of user A as the first-level lower-level users, users B and C are selected, user A is set as the top-level node, users B and C are set as the first-level lower-level nodes, and upper and lower relationship edges are formed between A and B and A and C. Then similarly, a number of users are randomly selected from the parents, relatives and friends of users B and C as the second-level lower-level users, which are set as the second-level lower-level nodes, and upper and lower relationship edges are formed between users B, C and their corresponding lower-level nodes, thereby constructing a three-level pyramid structure.

[0097] After the pyramid structure is completed, transaction data is generated based on the relationship between users at each level. First, the frequency and amount of transactions are determined. The amount of the transfer of the lower-level users is usually set to a small range, specifically between hundreds to thousands of yuan per transaction. The small amount of transfer is set to simulate the characteristics of the lower-level users frequently transferring funds to the middle-level users. This amount range can reflect the limitations of the economic capacity of the participants, while reducing the suspiciousness of a single transaction. Next, transaction records are generated, with the lower-level users initiating high-frequency small-amount transfers to the middle-level users, and the middle-level users making large-amount transfers to the top-level users. This transfer mode reflects the characteristics of the funds being concentrated layer by layer. Finally, the transaction date and time are set, with the transaction time being reasonably distributed across different months to simulate continuous fund flow, thereby enhancing the authenticity and diversity of the data.

[0098] In addition, feature engineering is performed and corresponding evaluation indicators are set to evaluate the effectiveness of illegal fund-raising detection based on the generated data.

[0099] Among them, feature engineering includes the bank to which the account belongs, the balance, the type of card, the age of the cardholder, the occupation of the cardholder, the statistical information of the transaction amount (mean, maximum, minimum, variance), and the number of transactions. The label is whether the account is marked as blacklisted (marked as 1 for blacklisted and 0 for whitelisted). The positive and negative sample ratio is 1:10. The training set and test set are divided: 30% of the data is allocated to the test set, and the remaining 70% of the data is used to train the model.

[0100] The evaluation indicators are set as follows:

[0101] Precision (Precision): represents the proportion of actual positive samples among the predicted positive samples. High precision indicates that the model can effectively identify illegal fund-raising transactions and reduce false positives. The specific formula is:

[0102]

[0103] Recall (Recall): represents the proportion of model-predicted positive samples among actual positive samples. A higher recall rate can ensure that as many illegal fund-raising transactions as possible are identified, reducing false negatives. The specific formula is:

[0104]

[0105] F1 Score (F1-Score): is the harmonic mean of precision and recall, used to balance the model performance when the data is imbalanced. The specific formula is:

[0106]

[0107] KS value: is an important indicator to measure the discrimination ability of binary classification model, widely used in credit scoring and risk prediction fields. KS value compares the maximum difference between the cumulative distribution of positive examples (such as illegal financing transactions) and negative examples (such as normal transactions) to evaluate the discrimination effect of the model on two types of samples. Its calculation formula is as follows:

[0108] KS = max |TPR - FPR|

[0109] In the formula, TPR represents the proportion of correctly predicted positive examples in all actual positive examples, which measures the recognition ability of the model to positive sample. FPR represents the proportion of false positive examples in all actual negative examples. It measures the possibility of misjudging negative samples as positive examples. The value range of KS value is between 0 and 1, and the closer the result is to 1, the better the performance of the model.

[0110] After generating the simulated transaction data (including normal transaction data and various types of illegal financing transaction data) by the above method, three existing models are used to detect illegal financing of the generated transaction data, and the above evaluation indicators are used to evaluate the detection effect, and the results are shown in Table 3 as follows.

[0111] Table 3 Comparison of illegal financing detection effects of different models

[0112]

[0113] From Table 3, it can be seen that the KS value of the existing three models in the illegal financing detection based on simulated transaction data is only about 60%, which shows that the generated transaction data successfully simulates the complex characteristics of illegal financing transactions, so that the existing model cannot achieve perfect classification and recognition effect, thus promoting the development of financial fraud detection and related research.

[0114] Effects of the embodiment

[0115] The financial data generation method for illegal financing transaction risk scenarios provided by the embodiment has the following beneficial effects compared with the existing financial data generation method:

[0116] Simulate real transaction data generation: by constructing a transaction data generation system based on user basic data, statistical data and relationship network, transaction data covering various illegal financing transaction modes and normal user transactions can be generated, covering detailed transaction modes of users of different professions and age groups.

[0117] Enhanced detection capability: For pyramid illegal fund-raising transaction scenarios, the generated corresponding illegal fund-raising transaction data has characteristics such as periodic transfer, pyramid structure and complex network relationship. Using such synthetic data can help improve the detection and analysis effect of illegal fund-raising behavior.

[0118] Data security and privacy protection: In the process of synthesizing data, the corresponding data is automatically generated according to the statistical law of each data, without involving real user data, ensuring data privacy and security, complying with legal and privacy protection requirements, and the synthetic data has characteristics very close to real user data, solving the problem of lack of real user data and difficulty in carrying out related research in the prior art.

[0119] Flexibility and adaptability: The method can flexibly adjust the transaction generation rules according to the characteristics of different professional groups and age groups of users, ensure that the generated data has high authenticity and diversity, and can be well applied to a variety of different application scenarios.

[0120] The above embodiments are only used to illustrate the specific implementation of the present application, and the present application is not limited to the description range of the above embodiments. Those skilled in the art should understand that the present application is not limited by the above embodiments, and the above embodiments and the description in the specification are only to illustrate the principle of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A financial data generation method for illegal fund-raising transaction risk scenarios, characterized by, The method comprises the following steps: Step S1, generating user base data of a plurality of users; Step S2, constructing a user relationship network of the plurality of users based on the user base data; Step S3, generating normal transaction data between the users based on the user base information, the user relationship network, statistical data, and predetermined normal transaction generation rules; Step S4, generating illegal fund-raising transaction data of a plurality of illegal fund-raising transaction scenarios between the users based on the user base information, the user relationship network, statistical data, and predetermined illegal fund-raising transaction generation rules, wherein the user base data of each user comprises bank card information and user information, and the user information at least contains age, income level, and occupation group, which are respectively generated according to corresponding statistical data, the user relationship network contains a plurality of nodes and edges between the nodes, wherein the nodes are the users, and the edges are the relationships between the users, in step S3, performing statistical analysis based on the user base data of the plurality of users to obtain feature data; determining a user group to which the plurality of users belong according to the feature data; generating the normal transaction generation rules of the user group according to the determined user group and a plurality of transaction data characteristic information corresponding to the user group; and generating the normal transaction data between the plurality of users based on the user base data and the normal transaction generation rules, in step S4, selecting a plurality of users with relationship edges from the user relationship network and constructing an illegal fund-raising relationship network based on the operation mode under the illegal fund-raising transaction scenario; generating the illegal fund-raising transaction generation rules based on transaction data characteristic information under the illegal fund-raising transaction scenario; and generating the illegal fund-raising transaction data between the plurality of users based on the user base data of the users in the illegal fund-raising relationship network and the illegal fund-raising transaction generation rules, the illegal fund-raising transaction scenario is a pyramid selling illegal fund-raising transaction scenario, in step S4, randomly selecting a user from the user relationship network as a top-level user according to the operation mode specific to the pyramid selling illegal fund-raising transaction scenario, selecting one or more from the friends, relatives, and parents of the top-level user as first-level subordinate users of the top-level user, and forming a superior-subordinate relationship edge between the top-level user and the first-level subordinate users, selecting one or more from the friends, relatives, and parents of one or more first-level subordinate users as second-level subordinate users, and forming a superior-subordinate relationship edge between the first-level subordinate users and the corresponding second-level subordinate users, repeating the above process to form a multi-level pyramid illegal fund-raising relationship network.

2. The financial data generation method for illegal fund-raising transaction risk scenarios according to claim 1, characterized in that: wherein the bank card information comprises a bank card number, a bank to which the bank card belongs, and a balance, the user information further comprises a name, a registration date, a gender, and a suspicious object marker, In step S1, for the bank card information, the bank card number is generated by a unique identifier, and the balance is randomly set within a predetermined numerical range according to the corresponding income level, For the user information, the income level and the occupation group are allocated based on the corresponding statistical data according to the proportion of age distribution.

3. The financial data generation method for illegal fund-raising transaction risk scenarios according to claim 1, characterized in that: wherein The user relationship network includes multiple nodes and edges between the nodes, the nodes being the users, and the edges being the relationships between the users, The types of relationships include parent-child relationships, relative relationships, and friend relationships, the friend relationships including generally related friend relationships and better related friend relationships, In step S2, an initial user relationship network is generated based on the user basic data and the types of relationships and is dynamically updated.

4. The financial data generation method for illegal fund-raising transaction risk scenarios according to claim 3, characterized in that: wherein In step S2, for the friend relationships, according to the age difference between a target user in the users and the users, a potential friend group of the target user is determined; two groups of a predetermined number of users in the potential friend group are randomly selected and set as the target user's generally related friends and better related friends, respectively, and the bidirectionality of the friend relationships is ensured, and a predetermined number of users selected from the friends of the target user's friends are also set as the target user's friends, For the relative relationships, a predetermined number of users are randomly selected for the target user and set as the target user's relatives, and the bidirectionality of the relative relationships is ensured, For the parent-child relationships, according to the age difference between the target user and the users and a predetermined logical rule, the users are selected for the target user and set as the target user's parents or children, and the uniqueness of the parent-child relationships is ensured.

5. The financial data generation method for illegal fund-raising transaction risk scenarios according to claim 1, characterized in that: wherein The transaction data characteristic information includes transaction object range, transaction time range, transaction amount range, transaction frequency range, and transaction intention range, which are set according to the age, income level, and occupation group of the users in the user group.

6. The financial data generation method for illegal fund-raising transaction risk scenarios according to claim 1, characterized in that: wherein The transaction data characteristic information includes transaction object range, transaction time range, transaction amount range, and transaction frequency range, which are set according to the operation mode of the pyramid scheme illegal fund-raising transaction scenario, The transaction object range includes the user's superior users and subordinate users, and the proportion range of the number of payment and collection transactions between the user and the superior users and between the user and the subordinate users is set, respectively, The transaction frequency range is one to several times per week or per month in a periodic manner.

7. The financial data generation method for illegal fund-raising transaction risk scenarios according to claim 1, characterized in that: wherein The illegal fund-raising transaction scenarios include a pyramid selling type illegal fund-raising transaction scenario, a virtual currency illegal fund-raising transaction scenario, a real estate illegal fund-raising transaction scenario, an old-age project illegal fund-raising transaction scenario and an internet financial platform illegal fund-raising transaction scenario.

Images