Transaction early warning method and device for anonymous virtual wallet

Abstract

The invention discloses a transaction early warning method and device for an anonymous virtual wallet, and relates to the field of financial science and technology or other related technical fields, and the method comprises the steps: collecting transaction data of a target financial transaction through the anonymous virtual wallet, and carrying out the feature extraction and preprocessing of the transaction data; inputting the preprocessed transaction data into an illegal fund transfer early warning model, and calculating the posterior probability of the transaction data belonging to each category by the illegal fund transfer early warning model; and according to the calculated posterior probability, selecting the category with the highest posterior probability as a prediction result, and outputting alarm information for the target financial transaction under the condition that the prediction result indicates that the target financial transaction relates to illegal fund transfer. The technical problem that illegal fund transactions performed by an anonymous virtual wallet cannot be identified by using a rule model in related technologies is solved.

Description

Technical Field

[0001] The present invention relates to the field of financial technology or other related fields, and in particular to a transaction early warning method and device for an anonymous virtual wallet. Background Art

[0002] In recent years, with the complexity of the global financial environment, the identification of illegal fund transfers has faced many challenges, and the compliance requirements for financial institutions have become increasingly stringent. In the relevant technologies, the illegal fund transfer identification methods adopted by financial institutions mainly rely on manual analysis and identification of suspicious transactions, and the suspicious transaction warning mechanism adopted mainly relies on the basic information and transaction data of customers, including but not limited to the nationality, address, age, occupation of customers, and the specific amount and purpose of transactions. This information is used to build a rule model, and the warning indicators are determined through repeated training. However, this rule-based model training and updating method has limitations: the accuracy of the model is highly dependent on the personal information provided by the customer. In scenarios where customer information is incomplete or anonymous transactions are conducted, such as transactions conducted through anonymous virtual wallets, the prediction effect of the model will be greatly reduced. At the same time, the existing rule model is mainly based on traditional transaction features, and it is difficult to effectively identify and warn of illegal fund transfers in anonymous virtual wallet transactions.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present invention provide a transaction warning method and device for an anonymous virtual wallet, so as to at least solve the technical problem in the related art that the rule model cannot identify illegal fund transactions performed by anonymous virtual wallets.

[0005] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a transaction warning method for an anonymous virtual wallet is provided, comprising: collecting transaction data of a target financial transaction using an anonymous virtual wallet, and performing feature extraction and preprocessing on the transaction data; inputting the preprocessed transaction data into an illegal funds transfer warning model, and using the illegal funds transfer warning model to calculate the posterior probability that the transaction data belongs to each category; based on the calculated posterior probabilities, selecting the category with the highest posterior probability as the prediction result, wherein, when the prediction result indicates that the target financial transaction involves illegal funds transfer, outputting warning information for the target financial transaction.

[0006] Optionally, the steps of collecting transaction data of target financial transactions using an anonymous virtual wallet, and extracting and preprocessing features of the transaction data include: collecting transaction data when a consumer uses the anonymous virtual wallet to conduct a target financial transaction, wherein the transaction data includes at least one of the following: transaction time period, transaction amount, wallet balance after the transaction, transaction currency, transaction type, whether cross-border transactions are involved, and transaction purpose; extracting a feature set associated with an illegal funds transfer warning from the transaction data, wherein the feature set includes: transaction time period, transaction amount, wallet balance, currency type, loan identification, transaction type, and cross-border transaction flag; preprocessing the extracted feature set; and outputting the preprocessed transaction data as input to the illegal funds transfer warning model.

[0007] Optionally, the step of preprocessing the extracted feature set includes: performing outlier detection on the transaction amount feature in the feature set, converting abnormal amount values ​​that are negative or greater than a first preset abnormal threshold to zero; or, truncating abnormal amount values ​​that are negative or greater than a first preset abnormal threshold; detecting whether there are abnormal values ​​in the wallet balance feature in the feature set, setting an abnormal value greater than the maximum wallet amount as the maximum boundary amount value, or setting an abnormal value less than the minimum wallet amount to the minimum boundary amount value; checking whether the transaction time period in the feature set exceeds the normal range, and setting the abnormal transaction time period to a specified time value.

[0008] Optionally, the step of preprocessing the extracted feature set includes: for numerical features in the feature set, using a normalization formula to standardize the feature values, wherein the numerical features include: transaction time period, transaction amount and wallet balance; for categorical features in the feature set, label encoding is performed, and the feature value of each feature is converted into an integer code, wherein the categorical features include: currency type, loan identifier, transaction type and cross-border transaction flag.

[0009] Optionally, the illegal funds transfer early warning model is trained in the following manner: feature extraction and preprocessing of historical transaction data in the historical process; using the preprocessed transaction data set to train a Bayesian classification model, the Bayesian classification model calculates the prior probabilities of positive categories and negative categories, and the Bayesian classification model calculates the conditional probability of each feature in each category, wherein the positive category is used to indicate data involving illegal funds transfer in the transaction data set, and the negative category is used to indicate data not involving illegal funds transfer in the transaction data set; according to the training results, the model parameters of the Bayesian classification model are updated until the accuracy of the posterior probability of the transaction data output by the Bayesian classification model belonging to each category reaches a preset accuracy threshold, and the model training is determined to be completed to obtain the illegal funds transfer early warning model.

[0010] Optionally, it also includes: re-extracting features and preprocessing data on the new transaction data according to the newly collected transaction data, and dynamically adjusting model parameters of the illegal funds transfer early warning model based on the preprocessed new transaction data.

[0011] Optionally, the step of selecting a category with the highest posterior probability as the prediction result based on the calculated posterior probability includes: comparing the calculated posterior probabilities of the positive category and the negative category; selecting the category with the largest posterior probability value, and using the selected category as the prediction result; when the posterior probability of the positive category is greater than that of the negative category, determining that the target financial transaction involves illegal fund transfer; when the posterior probability of the negative category is greater than that of the positive category, determining that the target financial transaction does not involve illegal fund transfer.

[0012] According to another aspect of an embodiment of the present invention, a transaction warning device for an anonymous virtual wallet is provided, comprising: a transaction data collection unit, for collecting transaction data of a target financial transaction using an anonymous virtual wallet, and performing feature extraction and preprocessing on the transaction data; a data input unit, for inputting the preprocessed transaction data into an illegal funds transfer warning model, and the illegal funds transfer warning model calculates the posterior probability that the transaction data belongs to each category; a transaction warning unit, for selecting a category with a highest posterior probability as a prediction result based on the calculated posterior probability, wherein, when the prediction result indicates that the target financial transaction involves illegal funds transfer, outputting warning information for the target financial transaction.

[0013] Optionally, the transaction data collection unit includes: a transaction data collection module, used to collect transaction data when the consumer uses the anonymous virtual wallet to perform a target financial transaction, wherein the transaction data includes at least one of the following: transaction time period, transaction amount, wallet balance after the transaction, transaction currency, transaction type, whether cross-border transactions are involved, and transaction purpose; a feature extraction module, used to extract a feature set associated with an illegal funds transfer warning from the transaction data, wherein the feature set includes: transaction time period, transaction amount, wallet balance, currency type, loan identification, transaction type, and cross-border transaction flag; a feature preprocessing module, used to preprocess the extracted feature set; and a transaction data output module, used to output the preprocessed transaction data as input to the illegal funds transfer warning model.

[0014] Optionally, the feature preprocessing module includes: a first outlier detection submodule, used to perform outlier detection on the transaction amount feature in the feature set, and convert abnormal amount values ​​that are negative or greater than a first preset abnormal threshold to zero; or, truncate abnormal amount values ​​that are negative or greater than the first preset abnormal threshold; a second outlier detection submodule, used to detect whether there are abnormal values ​​in the wallet balance feature in the feature set, and set abnormal values ​​greater than the maximum wallet amount as the maximum boundary amount value, or, set abnormal values ​​less than the minimum wallet amount to the minimum boundary amount value; a time period detection submodule, used to check whether the transaction time period in the feature set exceeds the normal range, and set the abnormal transaction time period to a specified time value.

[0015] Optionally, the feature preprocessing module includes: a standardization submodule, used to use a normalization formula to standardize the feature values ​​of the numerical features in the feature set, wherein the numerical features include: transaction time period, transaction amount and wallet balance; an encoding submodule, used to perform label encoding on the categorical features in the feature set, and convert the feature value of each feature into an integer code, wherein the categorical features include: currency type, loan identifier, transaction type and cross-border transaction flag.

[0016] Optionally, the illegal funds transfer early warning model is obtained by training the following units: a feature extraction subunit, used to extract features and preprocess historical transaction data in the historical process; a training subunit, used to train a Bayesian classification model using the preprocessed transaction data set, the Bayesian classification model calculates the prior probabilities of positive categories and negative categories, and the Bayesian classification model calculates the conditional probability of each feature under each category, wherein the positive category is used to indicate data involving illegal funds transfer in the transaction data set, and the negative category is used to indicate data not involving illegal funds transfer in the transaction data set; a parameter updating subunit, used to update the model parameters of the Bayesian classification model according to the training results, until the accuracy of the posterior probability of the transaction data output by the Bayesian classification model belonging to each category reaches a preset accuracy threshold, it is determined that the model training is completed, and the illegal funds transfer early warning model is obtained.

[0017] Optionally, the transaction early warning device for an anonymous virtual wallet further includes: a feature re-extraction unit, which is used to re-extract features and pre-process data on new transaction data according to newly collected transaction data, and dynamically adjust model parameters of the illegal funds transfer early warning model based on the pre-processed new transaction data.

[0018] Optionally, the transaction warning unit includes: a probability comparison module, used to compare the calculated posterior probabilities of the positive category and the negative category; a category selection module, used to select the category with the largest posterior probability value, and use the selected category as the prediction result; when the posterior probability of the positive category is greater than that of the negative category, it is determined that the target financial transaction is suspected of illegal fund transfer; when the posterior probability of the negative category is greater than that of the positive category, it is determined that the target financial transaction does not involve illegal fund transfer.

[0019] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned transaction warning methods for an anonymous virtual wallet.

[0020] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the above-mentioned transaction warning methods for an anonymous virtual wallet.

[0021] According to another aspect of an embodiment of the present invention, a computer program product is further provided, including a computer program, wherein when the computer program is executed by a processor, the steps of any one of the above-mentioned transaction warning methods for an anonymous virtual wallet are implemented.

[0022] In the present disclosure, transaction data of target financial transactions using anonymous virtual wallets are collected, and feature extraction and preprocessing are performed on the transaction data; the preprocessed transaction data are input into an illegal funds transfer early warning model, and the illegal funds transfer early warning model calculates the posterior probability that the transaction data belongs to each category; based on the calculated posterior probabilities, the category with the highest posterior probability is selected as the prediction result, wherein, when the prediction result indicates that the target financial transaction involves illegal funds transfer, an alarm message for the target financial transaction is output.

[0023] In the present disclosure, in the absence of customer information, by monitoring the transaction data of customers who use anonymous virtual wallets, and then using an early warning model to classify the transactions and automatically iterate the model, dynamic suspicious transaction early warnings are performed, thereby accurately identifying information on illegal fund transfers using anonymous virtual wallets, and providing early warning information in a timely manner, thereby solving the technical problem in related technologies that the rule model cannot identify illegal fund transactions conducted by anonymous virtual wallets. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0025] Figure 1 A hardware structure block diagram of a computer terminal (or mobile device) for implementing a transaction early warning method for an anonymous virtual wallet is shown;

[0026] Figure 2 is a flow chart of an optional transaction warning method for an anonymous virtual wallet according to an embodiment of the present invention;

[0027] Figure 3 is a schematic diagram of an optional transaction warning device for an anonymous virtual wallet according to an embodiment of the present invention;

[0028] Figure 4 It is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0029] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] To facilitate those skilled in the art to understand the present invention, some terms or nouns involved in the embodiments of the present invention are explained below:

[0032] The Bayesian model is a statistical modeling method based on probability theory. Based on the Bayesian theorem, it provides a framework for calculating the posterior probability (i.e., the probability after considering the current data) by knowing the prior probability (i.e., the probability without considering the current data) and the likelihood probability (i.e., the probability of a hypothesis occurring when the data is given). In this application, in the classification of customers who identify illegal fund transfers, the Bayesian model is used to evaluate the probability that the transaction data belongs to the category of suspected illegal fund transfer or not suspected illegal fund transfer.

[0033] An anonymous virtual wallet is a tool used to store and transfer virtual funds in virtual transactions. It is designed to protect user privacy and allow users to conduct transactions without revealing their identity.

[0034] It should be noted that the transaction warning method and device for anonymous virtual wallets in the present disclosure can be used in the field of financial technology. When the detection of illegal fund transfers by customers using anonymous virtual wallets is realized based on the Bayesian model, it can also be used in any field other than the field of financial technology. When the detection of illegal fund transfers by customers using anonymous virtual wallets is realized based on the Bayesian model, the application field of the transaction warning method and device for anonymous virtual wallets in the present disclosure is not limited.

[0035] It should be noted that the information (including but not limited to user device information, virtual wallet information, user personal information, etc.) and data (including but not limited to transaction data for analysis, stored data, displayed data, etc.) collected by this disclosure are information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with the relevant laws, regulations and standards of the relevant regions, necessary confidentiality measures are taken, and public order and good customs are not violated, and corresponding operation portals are provided for users to choose to authorize or refuse. For example, an interface is set up between this system and relevant users or institutions. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or institution through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or institution.

[0036] It should be noted that in the present disclosure, when collecting and analyzing customer information, corresponding operation entrances are provided for users to choose to agree or reject the automated decision results; if the user chooses to reject, the expert decision-making process is entered.

[0037] The following embodiments of the present invention can be applied to various systems / applications / devices for transaction warning of anonymous virtual wallets. The present invention can overcome the defects of missing customer information and poor scalability of rule models. Based on the mobile terminal, location, and transaction behavior (including transaction time period, transaction amount, balance, transaction type, transaction currency, etc.) of the anonymous virtual wallet customer, the artificial intelligence machine learning technology is used to model, classify transactions and automatically iterate the model, and dynamically issue suspicious transaction warnings.

[0038] The present invention is described in detail below in conjunction with various embodiments.

[0039] Embodiment 1

[0040] According to an embodiment of the present invention, an embodiment of a transaction warning method for an anonymous virtual wallet is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0041] The transaction warning method embodiment for an anonymous virtual wallet provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 The hardware structure block diagram of a computer terminal (or mobile device) for implementing a transaction warning method for an anonymous virtual wallet is shown. Figure 1 As shown, the computer terminal 10 (or mobile device) may include one or more ( Figure 1The computer system includes a processor 102 (the processor 102 may include but is not limited to a microcontroller unit (MCU) or a field programmable gate array (FPGA)), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, the computer system may include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS), a network interface, a power supply and / or a camera. A person skilled in the art can understand that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0042] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or mobile device). As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0043] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the transaction warning method for an anonymous virtual wallet in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the transaction warning method for an anonymous virtual wallet is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0044] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0045] The display may be, for example, a touch screen liquid crystal display (LCD), which enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0046] Under the above operating environment, this application provides Figure 2 The transaction warning method for an anonymous virtual wallet is shown. Figure 2 The present invention is an optional flowchart of a transaction warning method for an anonymous virtual wallet according to an embodiment of the present invention.

[0047] It should be noted that the transaction warning method for an anonymous virtual wallet provided in this embodiment is applied to monitoring scenarios for identifying the risk of illegal fund transfer in financial transactions using anonymous virtual wallets. When a user chooses an anonymous wallet to conduct a transaction, this embodiment can warn of potential illegal fund transfer activities by analyzing the characteristics of transaction data, such as transaction time period, amount, balance, transaction type, currency, and location information, and can still ensure the accuracy of the warning even when customer information is missing.

[0048] like Figure 2 As shown, the transaction warning method for an anonymous virtual wallet includes the following steps:

[0049] Step S101, collecting transaction data of target financial transactions using anonymous virtual wallets, and performing feature extraction and preprocessing on the transaction data.

[0050] Among them, step S101 of this embodiment may include: collecting transaction data when a consumer uses an anonymous virtual wallet to conduct a target financial transaction, wherein the transaction data includes at least one of the following: transaction time period, transaction amount, wallet balance after the transaction, transaction currency, transaction type, whether cross-border transactions are involved, and transaction purpose; extracting a feature set associated with illegal funds transfer warning from the transaction data, wherein the feature set includes: transaction time period, transaction amount, wallet balance, currency type, loan identification, transaction type, and cross-border transaction flag; preprocessing the extracted feature set; and outputting the preprocessed transaction data as input to the illegal funds transfer warning model.

[0051] This step collects data records generated when consumers use anonymous virtual wallets for financial transactions. The collected transaction data covers multiple dimensions, including but not limited to the transaction time period, transaction amount, wallet balance after the transaction, transaction currency, transaction type (for example, transfer, payment, remittance, etc.), whether it involves cross-border transactions, and the purpose of the transaction (if available). Due to the characteristics of anonymous virtual wallets, the transaction data may not contain traditional customer identity information, such as name, address, occupation, etc. Therefore, this embodiment focuses on collecting data related to transaction behavior to make up for the challenges brought by missing information.

[0052] By screening out key features directly related to the early warning of illegal fund transfer from the collected transaction data, a feature set is constructed. The feature set includes transaction time period, transaction amount, wallet balance, currency type, loan identification (i.e. transaction direction, whether it is income or expenditure), transaction type and cross-border transaction mark. These feature information can reflect the time pattern of transactions, amount scale, direction of fund flow and traces of international transactions, and is an important basis for identifying illegal fund transfers.

[0053] After feature extraction, the data needs to be cleaned, standardized, and encoded to meet the requirements of model input and improve the efficiency of model training and prediction. Among them, cleaning processing can refer to removing or correcting errors and outliers in transaction data, such as unreasonable transaction amounts, timestamp errors, etc.; standardization processing refers to standardizing numerical features such as transaction amounts and wallet balances to ensure that features in different numerical ranges are comparable in the model. Coding processing refers to encoding categorical features, such as currency type, transaction type, and cross-border transaction signs, and converting them into numerical forms readable by the model. The preprocessed transaction data is used as the input of the illegal funds transfer warning model. After being sorted, these data can be directly used for model training or real-time prediction.

[0054] It should be noted that the transaction data of this embodiment should be organized into a format that can be directly processed by the model, usually including feature vectors and corresponding labels (whether illegal fund transfer is involved). The output data ensures the efficiency of model training and the accuracy of prediction results, and can achieve effective early warning of illegal fund transfer even in the use scenario of anonymous virtual wallets.

[0055] For the above steps: preprocessing the extracted feature set may include: performing outlier detection on the transaction amount feature in the feature set, converting abnormal amount values ​​that are negative or greater than a first preset abnormal threshold to zero; or, truncating abnormal amount values ​​that are negative or greater than the first preset abnormal threshold; detecting whether there are abnormal values ​​in the wallet balance feature in the feature set, setting the abnormal value greater than the maximum wallet amount as the maximum boundary amount value, or setting the abnormal value less than the minimum wallet amount to the minimum boundary amount value; checking whether the transaction time period in the feature set exceeds the normal range, and setting the abnormal transaction time period to a specified time value.

[0056] In this embodiment, when preprocessing the feature set extracted from the transaction data, it is first necessary to detect and process the transaction amount outliers, that is, to identify and process the outliers in the transaction amount to avoid the influence of extreme data during model training or prediction. Here, the transaction amount feature is mainly detected for outliers, and the detection range includes negative values ​​and values ​​exceeding a specific threshold. If the transaction amount feature contains a negative value or the value is greater than the set first preset abnormal threshold, one of the following processing measures is adopted: the negative value or the abnormally large value is directly set to zero (0) to eliminate their influence on the model; the negative value or the abnormally large value is truncated, that is, the value size is limited to ensure that the model prediction result is not distorted due to extreme values.

[0057] It should also be noted that, when preprocessing the feature set extracted from the transaction data, this embodiment also needs to detect and process wallet balance feature outliers to ensure that the wallet balance feature reflects the actual reasonable funds holding situation. Here, the wallet balance feature is mainly detected to see if there are outliers beyond the normal range. When it is detected that the value in the wallet balance feature is greater than the maximum limit of the wallet, it is set to the maximum boundary limit value to avoid the model's misunderstanding of abnormally large amounts of funds. Similarly, if the value is less than the minimum limit of the wallet, it is set to the minimum boundary limit value to ensure the rationality of the model input data.

[0058] Furthermore, when preprocessing the feature set extracted from the transaction data, this embodiment also needs to detect and process the abnormal value of the transaction time period to normalize the transaction time period data and ensure that all transaction time marks are within an acceptable normal range. Here, the transaction time period recorded in the transaction data is mainly checked to determine whether it exceeds the normal time range of a day (usually 0 to 24 hours). If it is found that the transaction time period record exceeds the normal time range, it is set to a specified reasonable time value, such as setting the abnormal time uniformly to 24 hours, so as to maintain the continuity and integrity of the data.

[0059] Optionally, the step of preprocessing the extracted feature set includes: for numerical features in the feature set, using a normalization formula to standardize the feature values, wherein the numerical features include: transaction time period, transaction amount and wallet balance; for categorical features in the feature set, label encoding is performed, and the feature value of each feature is converted into an integer code, wherein the categorical features include: currency type, loan identifier, transaction type and cross-border transaction flag.

[0060] Here, the purpose of normalizing numerical features is to map values ​​of different ranges to a unified scale to avoid model learning bias caused by differences in numerical ranges between features. Here, a normalization formula is needed to normalize the numerical features. The normalization formula used can be:

[0061]

[0062] Among them, x is the original eigenvalue, x min With x max are the minimum and maximum values ​​of the feature in the data set, respectively, normzlized is the normalized feature value. In this application, numerical features include transaction time period, transaction amount, wallet balance, etc. These features are standardized by the above formula and mapped to the range of 0 to 1 to ensure that they are comparable during model training and prediction.

[0063] The encoding of categorical features is to convert non-numeric information into numerical form for processing by machine learning models. Label encoding can be used here to convert the feature value of each feature into an integer encoding, that is, to map different category values ​​into a series of different integers. In this embodiment, categorical features include currency type, loan identifier (representing the transaction direction, such as borrowing or lending), transaction type (such as transfer, payment, etc.) and cross-border transaction flags. Through label encoding, these feature values ​​are converted into integers that increase from 0. For example, if there are three types of currency, they can be encoded as 0, 1, and 2, and the loan identifier can be encoded as 0 for borrowing, 1 for lending, and so on.

[0064] Step S102, inputting the pre-processed transaction data into an illegal funds transfer early warning model, and using the illegal funds transfer early warning model to calculate the posterior probability that the transaction data belongs to each category.

[0065] It should be noted in advance that the illegal funds transfer warning model mentioned in this embodiment is trained in the following manner: feature extraction and preprocessing of historical transaction data in the historical process; using the preprocessed transaction data set to train the Bayesian classification model, the Bayesian classification model calculates the prior probabilities of positive categories and negative categories, and the Bayesian classification model calculates the conditional probability of each feature under each category, wherein the positive category is used to indicate data involving illegal funds transfer in the transaction data set, and the negative category is used to indicate data not involving illegal funds transfer in the transaction data set; according to the training results, the model parameters of the Bayesian classification model are updated until the accuracy of the posterior probability of the transaction data output by the Bayesian classification model belonging to each category reaches a preset accuracy threshold, and the model training is determined to be completed, and an illegal funds transfer warning model is obtained.

[0066] Before training the model, extract features related to illegal fund transfer warning from historical transaction data, and ensure the quality of feature data. It is necessary to collect historical transaction data in advance, which covers the target financial transaction information using anonymous virtual wallets, and then normalize the numerical features (such as transaction time period, transaction amount, and wallet balance) to ensure the consistency of the numerical range. Label encode categorical features (such as currency type, loan identification, transaction type, and cross-border transaction flag) to convert non-numerical information into a numerical form that can be processed by the machine learning model.

[0067] The Bayesian classification model is a probability-based machine learning model used for classification prediction problems. Its core is the Bayesian theorem and the feature independence assumption. In this embodiment, based on the preprocessed transaction data set, the model identifies transaction data suspected of illegal fund transfer as a positive category, and transaction data not involving illegal fund transfer as a negative category. The model first calculates the prior probability of positive and negative categories, that is, the probability that the transaction data belongs to suspected illegal fund transfer or not involving illegal fund transfer without considering the features. The model further calculates the conditional probability of each feature under each category, that is, the probability of a specific feature appearing when a certain category is given. Based on the calculated prior and conditional probabilities, the model parameters will be continuously adjusted and optimized to improve the accuracy of the prediction. During the training process, the accuracy of the posterior probability of the transaction data output by the monitoring model belonging to the positive category (suspected of illegal fund transfer) and the negative category (not involving illegal fund transfer) is monitored. This accuracy needs to reach a preset accuracy threshold to ensure that the prediction results of the model are sufficiently reliable. When the accuracy of the model reaches or exceeds the preset accuracy threshold, the model training is considered to be completed, and the model parameter configuration at this time can effectively identify illegal fund transfer activities.

[0068] The trained illegal funds transfer warning model can be applied to the real-time transaction monitoring system to analyze the newly occurred transaction data and automatically output the probability that the transaction is suspected of illegal funds transfer. As more real-time transaction data is collected, the model can be continuously updated and optimized to adapt to the ever-changing transaction patterns and illegal funds transfer methods, and maintain the latest and effectiveness of its warning effect. Optionally, this embodiment also includes: according to the newly collected transaction data, re-extracting features and preprocessing the new transaction data, and dynamically adjusting the model parameters of the illegal funds transfer warning model based on the preprocessed new transaction data.

[0069] When adjusting the model parameters, it is necessary to continuously collect new transaction data, which may come from the real-time transaction records of anonymous virtual wallets. Similar to the processing of historical data, the feature set associated with the warning of illegal fund transfer is extracted from the new transaction data, including transaction time period, transaction amount, wallet balance, currency type, loan identification, transaction type and cross-border transaction flag, etc. The newly extracted feature set is preprocessed, including the standardization of numerical features (such as the normalization of transaction amount and wallet balance) and the encoding of categorical features (such as the label encoding of currency type and transaction type) to ensure the consistency of the format and scale of the new data with the historical data. Based on the characteristics of the Bayesian classification model, the addition of new data can be regarded as an update to the model parameters. Specifically, the prior probabilities of the positive category (suspected illegal fund transfer transactions) and the negative category (normal transactions) are recalculated using the newly collected and preprocessed transaction data. At the same time, the conditional probability of each feature under each category is updated.

[0070] Through the above steps, the dynamic recalculation of conditional probability and prior probability ensures that the model can capture potential illegal fund transfer signals in transaction data in a timely manner. The feature set of new transaction data is input into the model as the basis for dynamic adjustment of the model. The model automatically adjusts its parameters by analyzing new data, including the update of prior probability and conditional probability. After the parameter adjustment, the model will perform self-verification and evaluate the prediction performance of the adjusted model. If the accuracy and stability of the model meet the preset standards, the adjusted parameters will be saved for subsequent analysis and prediction of transaction data.

[0071] By dynamically adjusting the model parameters, this application can enable the illegal funds transfer warning model to maintain efficient and accurate warning capabilities in a constantly changing trading environment. As the system continues to operate, the model will continue to evolve and improve its efficiency in identifying emerging illegal funds transfer methods.

[0072] Step S103, based on the calculated posterior probability, select the category with the highest posterior probability as the prediction result, wherein, when the prediction result indicates that the target financial transaction involves illegal fund transfer, output warning information for the target financial transaction to prevent the transaction from being completed.

[0073] It should be noted that step S103 in this embodiment includes: comparing the calculated posterior probabilities of the positive category and the negative category; selecting the category with the largest posterior probability value, and using the selected category as the prediction result; when the posterior probability of the positive category is greater than that of the negative category, determining that the target financial transaction is suspected of illegal fund transfer; when the posterior probability of the negative category is greater than that of the positive category, determining that the target financial transaction does not involve illegal fund transfer.

[0074] During the model training phase, the Bayesian classification model has learned the characteristics of transaction data and its association with positive categories (transactions suspected of illegal fund transfer) and negative categories (normal transactions). For each new target financial transaction data, the model will calculate the posterior probability of belonging to the positive category and the negative category based on its feature value. The posterior probability is the probability assessment of the transaction data belonging to a certain category after considering the transaction feature information, reflecting the possibility of the transaction data belonging to the positive category or the negative category under the given characteristics. After the model calculates the posterior probabilities of the positive category and the negative category, it will compare them to determine which category the transaction data is more likely to belong to, so as to determine whether the transaction is suspected of illegal fund transfer.

[0075] Specifically, during the comparison process, it is necessary to directly compare the posterior probability of the positive category with the posterior probability of the negative category to find the category with the larger probability value. Then, this embodiment uses the following judgment rules based on the comparison results to determine the final prediction result: If the posterior probability of the positive category is greater than the posterior probability of the negative category, the model determines that the target financial transaction is suspected of illegal fund transfer. At this time, the early warning system will mark the transaction and may trigger further investigation or block the transaction. If the posterior probability of the negative category is greater than the posterior probability of the positive category, the model determines that the target financial transaction does not involve illegal fund transfer. In this case, the transaction will be regarded as a normal financial activity and can continue.

[0076] Through the posterior probability calculated by the Bayesian model, this embodiment realizes the intelligent classification of transaction data, not only considering the independent influence of a single feature, but also integrating the joint influence of all features, thereby improving the accuracy of judging illegal fund transfers. At the same time, the probability-based judgment mechanism enables the model to express the suspicious degree of the transaction in a quantitative form, providing decision makers with a more objective and quantitative early warning basis.

[0077] This embodiment has the ability to automatically iterate and update, and can continuously adjust model parameters according to newly collected transaction data, ensuring that the accuracy and effectiveness of the early warning model are continuously optimized as the transaction scenario changes. This dynamic adjustment mechanism greatly reduces the human resource cost of model maintenance.

[0078] Through the above steps, the transaction data of the target financial transaction using the anonymous virtual wallet can be collected, the transaction data can be feature extracted and preprocessed, the preprocessed transaction data can be input into the illegal funds transfer warning model, the illegal funds transfer warning model can calculate the posterior probability that the transaction data belongs to each category, and according to the calculated posterior probability, the category with the highest posterior probability can be selected as the prediction result, wherein, when the prediction result indicates that the target financial transaction involves illegal funds transfer, the warning information for the target financial transaction can be output to prevent the transaction from being completed. In this embodiment, in the case of missing customer information, by monitoring the transaction data of customers using anonymous virtual wallets, and then using the warning model to classify the transactions and automatically iterate the model, a suspicious transaction warning can be dynamically issued, thereby accurately identifying the information of illegal funds transfer using anonymous virtual wallets, and providing warning information in a timely manner, thereby solving the technical problem in the related art that the rule model cannot identify illegal funds transactions conducted by anonymous virtual wallets.

[0079] In this embodiment, the characteristic patterns of anonymous virtual wallet transactions can be automatically identified and learned through machine learning methods, overcoming the problem that the warning effect of traditional rule models is greatly reduced when customer information is missing. Even in an anonymous environment, suspicious transactions can be accurately identified through transaction behavior characteristics, improving the intelligence level of the entire anti-illegal fund transfer warning system.

[0080] At the same time, this embodiment uses the posterior probability calculated by the Bayesian model to classify transactions, which can more accurately determine whether the target financial transaction is suspected of illegal fund transfer. Compared with the traditional identification scheme that only relies on basic customer information, it can capture more potential illegal fund transfer signals from transaction behaviors, thereby improving the accuracy of early warning.

[0081] Another optional specific implementation is described in detail below.

[0082] Step 1: Collect transaction data for training the model. The data set should include the following fields: transaction data, including: transaction time period, transaction amount, account balance, currency code, loan sign, transaction type and cross-border sign, as well as the label of each transaction, that is, whether it is suspected of illegal fund transfer. Among them, the transaction time period, transaction amount and account balance fields are numerical features, and the values ​​are positive real numbers; currency code, loan sign, transaction type and cross-border sign are categorical features, and the values ​​are natural numbers.

[0083] Step 2: Preprocess the transaction data, mainly including: (1) Data cleaning, cleaning the transaction time period, transaction amount and account balance features in the transaction data, and dealing with outliers. (2) Standardization, standardizing the numerical features to ensure that they are within a similar numerical range. (3) Label encoding, label encoding the categorical features.

[0084] Step 2 (1): A. Check whether the transaction amount feature has negative or abnormally large values. If so, set it to 0 or perform truncation. B. Check whether the account balance feature has abnormal values. If there is a value greater than the maximum amount of the account or less than the minimum amount of the account, set it to the corresponding boundary value. C. Transaction time period: Check whether the transaction time period exceeds the normal range (1 to 24 hours). If there is an abnormal value, set it to 24. D. Check whether the category feature has unknown or abnormal categories. If there is an abnormal value, delete the transaction data.

[0085] Step 2 (2): Process the numerical features according to the following normalization formula:

[0086]

[0087] Step 2 (3): Map the values ​​of the categorical features into integers that increase sequentially starting from 0.

[0088] Step 3: Train the Naive Bayes classifier model using the processed features and labels.

[0089] Step 3 (1): Assume that the eigenvector X = {x1, x2, x3...x n}. Calculate the prior probabilities of belonging to the positive category (suspected of illegal fund transfer) and the negative category (not suspected of illegal fund transfer) respectively.

[0090]

[0091] P(C0)=1-P(C1|X).

[0092] Step 3 (2): For each feature x i and each category C i , calculate the feature x under this category i The conditional probability P(C i |x i ). According to the Naive Bayes hypothesis, assuming that the features are independent of each other, then:

[0093]

[0094] Step 4: Prediction.

[0095] Step 4 (1): For new transaction data, its feature vector is Xτ={xτ1,xτ2,xτ3......x τn}, and calculate the posterior probability that it belongs to each category.

[0096]

[0097] Step 4 (2): Based on the calculated posterior probability, select the category with the highest posterior probability as the prediction result.

[0098]

[0099] Step 5: Automatic iterative update. Based on the incoming new data, steps 2 to 3 are repeated continuously to continuously improve the model parameters.

[0100] Compared with the existing methods, the embodiment of the present invention proposes a dynamic suspicious transaction early warning method, which overcomes the shortcomings of low update frequency and high maintenance cost of the early warning model in the existing methods.

[0101] Through the above implementation steps, analysis and early warning can be carried out immediately when new transaction data arrives, without the need for regular manual model training and updating. This real-time monitoring capability is crucial for responding quickly to illegal fund transfer activities, and can stop the flow of illegal funds in a timely manner, reducing the impact of illegal fund transfer risks on financial institutions and society.

[0102] At the same time, through the above implementation, the accuracy of suspicious transaction warning can be guaranteed in the case of missing customer information. At the same time, the model can be dynamically adjusted as the transaction scenario changes, which greatly reduces the human resource cost and improves the timeliness of the identification of illegal fund transfers.

[0103] The following is a detailed description in conjunction with another embodiment.

[0104] Embodiment 2

[0105] The transaction warning device for an anonymous virtual wallet provided in this embodiment includes multiple implementation units, each of which corresponds to each implementation step in the above-mentioned embodiment 1.

[0106] Figure 3 is a schematic diagram of an optional transaction warning device for an anonymous virtual wallet according to an embodiment of the present invention, such as Figure 3 As shown, the transaction warning device for an anonymous virtual wallet may include: a transaction data collection unit 31, a data input unit 32, and a transaction warning unit 33.

[0107] The transaction data collection unit 31 is used to collect transaction data of target financial transactions using anonymous virtual wallets, and perform feature extraction and preprocessing on the transaction data.

[0108] The data input unit 32 is used to input the pre-processed transaction data into the illegal funds transfer early warning model, and the illegal funds transfer early warning model calculates the posterior probability that the transaction data belongs to each category.

[0109] The transaction warning unit 33 is used to select the category with the highest posterior probability as the prediction result according to the calculated posterior probability, wherein, when the prediction result indicates that the target financial transaction involves illegal fund transfer, output warning information for the target financial transaction to prevent the transaction from being completed.

[0110] The transaction warning device for the anonymous virtual wallet can be used to collect transaction data of the target financial transaction using the anonymous virtual wallet through the transaction data collection unit 31, perform feature extraction and preprocessing on the transaction data, and then input the preprocessed transaction data into the illegal funds transfer warning model through the data input unit 32, and the illegal funds transfer warning model calculates the posterior probability that the transaction data belongs to each category, and then the transaction warning unit 33 selects the category with the highest posterior probability as the prediction result according to the calculated posterior probability, wherein, when the prediction result indicates that the target financial transaction involves illegal funds transfer, outputs warning information for the target financial transaction to prevent the transaction from being completed. In this embodiment, in the case of missing customer information, by monitoring the transaction data of customers using anonymous virtual wallets, and then using the warning model to classify the transactions and automatically iterate the model, a suspicious transaction warning is dynamically performed, so as to accurately identify the information of illegal funds transfer using anonymous virtual wallets, and provide warning information in a timely manner, thereby solving the technical problem in the related art that the rule model cannot identify the illegal funds transactions performed by anonymous virtual wallets.

[0111] Optionally, the transaction data collection unit includes: a transaction data collection module, which is used to collect transaction data when a consumer uses an anonymous virtual wallet to conduct a target financial transaction, wherein the transaction data includes at least one of the following: transaction time period, transaction amount, wallet balance after the transaction, transaction currency, transaction type, whether cross-border transactions are involved, and transaction purpose; a feature extraction module, which is used to extract a feature set associated with an illegal funds transfer warning from the transaction data, wherein the feature set includes: transaction time period, transaction amount, wallet balance, currency type, loan identification, transaction type, and cross-border transaction flag; a feature preprocessing module, which is used to preprocess the extracted feature set; and a transaction data output module, which is used to output the preprocessed transaction data as input to an illegal funds transfer warning model.

[0112] Optionally, the feature preprocessing module includes: a first outlier detection submodule, used to perform outlier detection on the transaction amount feature in the feature set, and convert the abnormal amount values ​​that are negative or greater than the first preset abnormal threshold to zero; or, truncate the abnormal amount values ​​that are negative or greater than the first preset abnormal threshold; a second outlier detection submodule, used to detect whether there are abnormal values ​​in the wallet balance feature in the feature set, and set the abnormal value greater than the maximum wallet amount as the maximum boundary amount value, or, set the abnormal value less than the minimum wallet amount to the minimum boundary amount value; a time period detection submodule, used to check whether the transaction time period in the feature set exceeds the normal range, and set the abnormal transaction time period to a specified time value.

[0113] Optionally, the feature preprocessing module includes: a standardization submodule, which is used to standardize the feature values ​​of the numerical features in the feature set using a normalization formula, wherein the numerical features include: transaction time period, transaction amount and wallet balance; an encoding submodule, which is used to perform label encoding on the categorical features in the feature set, and convert the feature value of each feature into an integer code, wherein the categorical features include: currency type, loan identifier, transaction type and cross-border transaction flag.

[0114] Optionally, the illegal funds transfer warning model is obtained by training the following units: a feature extraction subunit, used to extract features and preprocess historical transaction data in the historical process; a training subunit, used to train a Bayesian classification model using the preprocessed transaction data set, the Bayesian classification model calculates the prior probabilities of positive categories and negative categories, and the Bayesian classification model calculates the conditional probability of each feature under each category, wherein the positive category is used to indicate data involving illegal funds transfer in the transaction data set, and the negative category is used to indicate data not involving illegal funds transfer in the transaction data set; a parameter updating subunit, used to update the model parameters of the Bayesian classification model according to the training results, until the accuracy of the posterior probability of the transaction data output by the Bayesian classification model belonging to each category reaches a preset accuracy threshold, it is determined that the model training is completed, and the illegal funds transfer warning model is obtained.

[0115] Optionally, the transaction early warning device for an anonymous virtual wallet further includes: a feature re-extraction unit, which is used to re-extract features and pre-process the new transaction data according to the newly collected transaction data, and dynamically adjust the model parameters of the illegal funds transfer early warning model based on the pre-processed new transaction data.

[0116] Optionally, the transaction warning unit includes: a probability comparison module, used to compare the calculated posterior probabilities of the positive category and the negative category; a category selection module, used to select the category with the largest posterior probability value, and use the selected category as the prediction result; when the posterior probability of the positive category is greater than that of the negative category, it is determined that the target financial transaction is suspected of illegal fund transfer; when the posterior probability of the negative category is greater than that of the positive category, it is determined that the target financial transaction does not involve illegal fund transfer.

[0117] The transaction warning device for an anonymous virtual wallet may further include a processor and a memory. The transaction data collection unit 31, the data input unit 32, the transaction warning unit 33, etc. are all stored in the memory as program units, and the processor executes the program units stored in the memory to implement corresponding functions.

[0118] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set to identify financial transactions using anonymous virtual wallets for illegal fund transfers by adjusting kernel parameters.

[0119] The above-mentioned memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one storage chip.

[0120] Embodiment 3

[0121] An embodiment of the present application may provide an electronic device, Figure 4 is a structural block diagram of an electronic device according to an embodiment of the present application. Figure 4 As shown, the electronic device may include: one or more ( Figure 4 (only one is shown) processor 402, memory 404, storage controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0122] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the methods and devices in the embodiments of the present application, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0123] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: collect transaction data of the target financial transaction using the anonymous virtual wallet, and extract features and preprocess the transaction data; input the preprocessed transaction data into the illegal funds transfer early warning model, and the illegal funds transfer early warning model calculates the posterior probability that the transaction data belongs to each category; based on the calculated posterior probabilities, select the category with the highest posterior probability as the prediction result, wherein, when the prediction result indicates that the target financial transaction involves illegal funds transfer, output warning information for the target financial transaction.

[0124] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: collect transaction data when the consumer uses the anonymous virtual wallet to conduct a target financial transaction, wherein the transaction data includes at least one of the following: transaction time period, transaction amount, wallet balance after the transaction, transaction currency, transaction type, whether cross-border transactions are involved, and transaction purpose; extract a feature set associated with illegal fund transfer warning from the transaction data, wherein the feature set includes: transaction time period, transaction amount, wallet balance, currency type, loan identification, transaction type, and cross-border transaction flag; pre-process the extracted feature set; and output the pre-processed transaction data as input to the illegal fund transfer warning model.

[0125] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: perform abnormal value detection on the transaction amount feature in the feature set, and convert the abnormal amount value that is negative or greater than the first preset abnormal threshold to zero; or, truncate the abnormal amount value that is negative or greater than the first preset abnormal threshold; detect whether there are abnormal values ​​in the wallet balance feature in the feature set, and set the abnormal value greater than the maximum wallet amount as the maximum boundary amount value, or set the abnormal value less than the minimum wallet amount to the minimum boundary amount value; check whether the transaction time period in the feature set exceeds the normal range, and set the abnormal transaction time period to a specified time value.

[0126] The processor can call the information and application programs stored in the memory through the transmission device to perform the following steps: for the numerical features in the feature set, use the normalization formula to standardize the feature values, wherein the numerical features include: transaction time period, transaction amount and wallet balance; for the categorical features in the feature set, label encoding is performed, and the feature value of each feature is converted into an integer code, wherein the categorical features include: currency type, loan identification, transaction type and cross-border transaction flag.

[0127] The processor can call the information and application programs stored in the memory through the transmission device to perform the following steps: the illegal funds transfer warning model is trained in the following manner: feature extraction and preprocessing of historical transaction data in the historical process; using the preprocessed transaction data set to train the Bayesian classification model, the Bayesian classification model calculates the prior probabilities of positive and negative categories, and the Bayesian classification model calculates the conditional probability of each feature under each category, wherein the positive category is used to indicate data involving illegal funds transfer in the transaction data set, and the negative category is used to indicate data not involving illegal funds transfer in the transaction data set; according to the training results, the model parameters of the Bayesian classification model are updated until the accuracy of the posterior probability of the transaction data output by the Bayesian classification model belonging to each category reaches a preset accuracy threshold, and the model training is determined to be completed to obtain the illegal funds transfer warning model.

[0128] The processor can call the information and application programs stored in the memory through the transmission device to perform the following steps: re-extract features and pre-process the new transaction data based on the newly collected transaction data, and dynamically adjust the model parameters of the illegal funds transfer warning model based on the pre-processed new transaction data.

[0129] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: compare the calculated posterior probabilities of the positive category and the negative category; select the category with the largest posterior probability value, and use the selected category as the prediction result; when the posterior probability of the positive category is greater than that of the negative category, determine that the target financial transaction is suspected of illegal fund transfer; when the posterior probability of the negative category is greater than that of the positive category, determine that the target financial transaction does not involve illegal fund transfer.

[0130] It can be understood by those skilled in the art that Figure 4 The structure shown is for illustration only, and the electronic device may also be a smart phone, a tablet computer, a PDA, a mobile Internet device (MID), a PAD or other terminal device. Figure 4 The structure of the electronic device is not limited. Figure 4 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Figure 4 Different configurations shown.

[0131] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0132] Embodiment 4

[0133] The embodiment of the present application also provides a storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the transaction warning method for an anonymous virtual wallet provided in the first embodiment.

[0134] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, the computer-readable storage medium including a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the transaction warning method for an anonymous virtual wallet according to any one of the above-mentioned embodiments 1.

[0135] Optionally, in this embodiment, the above storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0136] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the transaction warning method for an anonymous virtual wallet described in each embodiment of the present application.

[0137] The present application also provides a computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the transaction warning method for an anonymous virtual wallet described in each embodiment of the present application are implemented.

[0138] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0139] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0140] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0141] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0142] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0143] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0144] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A transaction warning method for an anonymous virtual wallet, characterized in that: include: Collecting transaction data of target financial transactions using anonymous virtual wallets, and performing feature extraction and preprocessing on the transaction data; Inputting the preprocessed transaction data into an illegal funds transfer early warning model, and using the illegal funds transfer early warning model to calculate the posterior probability that the transaction data belongs to each category; According to the calculated posterior probability, the category with the highest posterior probability is selected as the prediction result, wherein, when the prediction result indicates that the target financial transaction involves illegal fund transfer, warning information for the target financial transaction is output.

2. The transaction early warning method according to claim 1, characterized in that: The steps of collecting transaction data of target financial transactions using an anonymous virtual wallet and performing feature extraction and preprocessing on the transaction data include: Collecting transaction data when the consumer uses the anonymous virtual wallet to conduct a target financial transaction, wherein the transaction data includes at least one of the following: transaction time period, transaction amount, wallet balance after the transaction, transaction currency, transaction type, whether cross-border transactions are involved, and transaction purpose; Extracting a feature set associated with an illegal funds transfer warning from the transaction data, wherein the feature set includes: transaction time period, transaction amount, wallet balance, currency type, loan identifier, transaction type, and cross-border transaction flag; Preprocessing the extracted feature set; The pre-processed transaction data is output as input to the illegal funds transfer early warning model.

3. The transaction early warning method according to claim 2, characterized in that: The step of preprocessing the extracted feature set comprises: Performing outlier detection on the transaction amount feature in the feature set, converting abnormal amount values ​​that are negative or greater than a first preset abnormal threshold to zero; or performing truncation processing on abnormal amount values ​​that are negative or greater than the first preset abnormal threshold; Detect whether there is an abnormal value in the wallet balance feature in the feature set, and set the abnormal value greater than the maximum amount of the wallet as the maximum boundary amount value, or set the abnormal value less than the minimum amount of the wallet as the minimum boundary amount value; Check whether the transaction time period in the feature set exceeds a normal range, and set the abnormal transaction time period to a specified time value.

4. The transaction early warning method according to claim 2, characterized in that: The step of preprocessing the extracted feature set comprises: For the numerical features in the feature set, a normalization formula is used to normalize the feature values, wherein the numerical features include: transaction time period, transaction amount and wallet balance; Label encoding is performed on the categorical features in the feature set, and the feature value of each feature is converted into an integer code, wherein the categorical features include: currency type, loan identifier, transaction type, and cross-border transaction flag.

5. The transaction early warning method according to claim 1, characterized in that: The illegal funds transfer early warning model is trained in the following way: Extract features and preprocess historical transaction data in the historical process; Using the preprocessed transaction data set to train a Bayesian classification model, the Bayesian classification model calculates the prior probabilities of positive categories and negative categories, and the Bayesian classification model calculates the conditional probability of each feature under each category, wherein the positive category is used to indicate data involving illegal fund transfer in the transaction data set, and the negative category is used to indicate data not involving illegal fund transfer in the transaction data set; According to the training results, the model parameters of the Bayesian classification model are updated until the accuracy of the posterior probability of the transaction data output by the Bayesian classification model belonging to each category reaches a preset accuracy threshold, and the model training is determined to be completed to obtain the illegal funds transfer early warning model.

6. The transaction early warning method according to claim 5, characterized in that: Also includes: According to the newly collected transaction data, feature extraction and data preprocessing are performed on the new transaction data again, and based on the preprocessed new transaction data, the model parameters of the illegal funds transfer early warning model are dynamically adjusted.

7. The transaction early warning method according to claim 1, characterized in that: According to the calculated posterior probability, the step of selecting the category with the highest posterior probability as the prediction result includes: Compare the calculated posterior probabilities of the positive and negative categories; Select the category with the largest posterior probability value and use the selected category as the prediction result; When the posterior probability of the positive class is greater than that of the negative class, determining that the target financial transaction is suspected of illegal fund transfer; When the posterior probability of the negative class is greater than that of the positive class, it is determined that the target financial transaction does not involve illegal fund transfer.

8. A transaction warning device for an anonymous virtual wallet, characterized in that: include: A transaction data collection unit, used to collect transaction data of target financial transactions using anonymous virtual wallets, and perform feature extraction and preprocessing on the transaction data; A data input unit, used for inputting the pre-processed transaction data into an illegal funds transfer early warning model, and the illegal funds transfer early warning model calculates the posterior probability that the transaction data belongs to each category; The transaction warning unit is used to select the category with the highest posterior probability as the prediction result according to the calculated posterior probability, wherein, when the prediction result indicates that the target financial transaction involves illegal fund transfer, output warning information for the target financial transaction.

9. An electronic device, characterized in that: The invention comprises one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the transaction warning method for an anonymous virtual wallet as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the transaction warning method for an anonymous virtual wallet described in any one of claims 1 to 7 are implemented.

Images