Blacklist generation and screening method, device, computer equipment and storage medium
By clustering and risk scoring of fund transaction records and blacklist screening combined with biometric information, the problems of misjudgment and misjudgment in traditional methods are solved, and the accuracy and accuracy of blacklist screening are improved.
Patent Information
- Application Number
- CN201811545308.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2018-12-17
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2038-12-17
AI Technical Summary
The screening methods of traditional blacklists in Internet finance are prone to misjudgment of "false positive" or "false negative", resulting in low screening accuracy and inability to accurately identify illegal and criminal activities.
By obtaining financial transaction records for clustering, risk categories are determined and scored using risk scoring models, it is divided into multi-level blacklists, and feature matching is combined with biometric information to improve screening accuracy.
It reduces the possibility of misjudgment and missed judgments, improves the accuracy of blacklist screening, and ensures accurate identification and early warning of illegal and criminal activities.
Smart Images

Figure CN109767076B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of Internet technology, and in particular to a method, device, computer equipment and storage medium for generating a blacklist and a method, device, computer equipment and storage medium for screening a blacklist. Background Art
[0002] With the development of Internet technology, Internet technology has gradually expanded into the financial field, forming a new business model - Internet finance. While Internet finance brings convenience, it also brings a series of problems. Some users use Internet finance to carry out some illegal and criminal activities, such as drug trafficking, money laundering, illegal fundraising, etc.
[0003] This has led to the creation of a blacklist punishment method, which uses blacklists to screen for illegal and criminal activities in Internet finance. In traditional methods, blacklist matching is mostly fuzzy matching, such as fuzzy matching of basic identity information such as name and ID number. This screening method is prone to "false positive" or "false negative" misjudgments, resulting in the inability to accurately screen out targets that hit the blacklist, and the accuracy of blacklist screening is low. Summary of the invention
[0004] Based on this, it is necessary to provide a blacklist generation method, device, computer equipment and storage medium and a blacklist screening method, device, computer equipment and storage medium that can improve the accuracy of blacklist screening in response to the above technical problems.
[0005] A method for generating a blacklist, the method comprising:
[0006] Obtain fund transaction records;
[0007] Clustering the fund accounts according to the fund transaction records;
[0008] Obtaining transaction features of each of the clusters;
[0009] Determining the risk category of the corresponding cluster according to the transaction characteristics;
[0010] Obtaining a risk scoring model corresponding to the risk category;
[0011] Scoring the fund accounts in the corresponding cluster using the risk scoring model to obtain a risk score;
[0012] Adding the fund account having the risk score greater than a first threshold to the first-level node of the blacklist;
[0013] The associated account of the capital account of the first-level node is obtained from the cluster, and the associated account is added to the second-level node of the blacklist.
[0014] In one embodiment, the method further comprises:
[0015] Obtaining the risk score of the fund account of the second-level node;
[0016] Comparing the risk score with a second threshold, and determining a target second-level node whose risk score is greater than the second threshold;
[0017] Obtain an associated account of the capital account corresponding to the target second-level node from the cluster, and add the associated account to a third-level node of the blacklist, where the third-level node is a third-level node corresponding to the target second-level node.
[0018] In one embodiment, the method further comprises:
[0019] Collecting biometric information of account holders of all fund accounts in the blacklist;
[0020] The biometric information is added to the corresponding node of the blacklist.
[0021] A blacklist screening method, the method comprising:
[0022] Receive business processing requests;
[0023] Extracting pending business information and biometric information of the requester from the business processing request;
[0024] Perform feature matching on the biometric information and the biometric information corresponding to the first-level node in the blacklist, and if there is a matching first-level node, obtain the risk category corresponding to the first-level node;
[0025] Calculating the matching degree between the pending business information and the transaction characteristics of the risk category, and if the matching degree is greater than a set value, determining that the business processing request hits the blacklist;
[0026] Generate an early warning task according to the business processing request.
[0027] In one of the embodiments, after the step of matching the biometric information with the biometric information corresponding to the first-level node in the blacklist, the method further includes:
[0028] If there is no matching first-level node, feature matching is performed between the biometric information and the biometric information corresponding to the second-level node in the blacklist;
[0029] If there is a matching second-level node, obtaining the biometric information of the first-level node corresponding to the matching second-level node;
[0030] Extracting biometric information of the transaction object from the business processing request;
[0031] Perform feature matching on the biometric information of the transaction object and the biometric information of the first-level node;
[0032] If there is a matching first-level node, it is determined that the service processing request hits the blacklist.
[0033] In one of the embodiments, after the step of matching the biometric information with the biometric information corresponding to the second-level node in the blacklist, the method further includes:
[0034] Acquire biometric information of the third-level node corresponding to the matched second-level node;
[0035] Perform feature matching on the biometric information of the transaction object and the biometric information corresponding to the third-level node, and if there is a matching third-level node, obtain the risk category corresponding to the third-level node;
[0036] The matching degree between the pending business information and the transaction characteristics of the risk category is calculated, and if the matching degree is greater than a set value, it is determined that the business processing request hits the blacklist.
[0037] A blacklist generation device, the device comprising:
[0038] Acquisition module, used to obtain fund transaction records;
[0039] A clustering module, used for clustering fund accounts according to the fund transaction records;
[0040] The acquisition module is also used to acquire transaction features of each cluster; determine the risk category of the corresponding cluster according to the transaction features; and acquire the risk scoring model corresponding to the risk category;
[0041] A scoring module, used to score the fund accounts in the corresponding cluster using the risk scoring model to obtain a risk score;
[0042] A blacklist generation module is used to add the capital account whose risk score is greater than a first threshold to the first-level node of the blacklist; obtain the associated account of the capital account of the first-level node from the cluster, and add the associated account to the second-level node of the blacklist.
[0043] A blacklist screening device, the device comprising:
[0044] A request receiving module, used to receive a business processing request;
[0045] An information extraction module, used to extract pending business information and biometric information of the requester from the business processing request;
[0046] A feature matching module, used to perform feature matching between the biometric feature information and the biometric feature information corresponding to the first-level node in the blacklist, and if there is a matching first-level node, obtain the risk category corresponding to the first-level node;
[0047] A determination module, used for calculating the matching degree between the pending business information and the transaction characteristics of the risk category, and if the matching degree is greater than a set value, determining that the business processing request hits the blacklist;
[0048] The task generation module is used to generate an early warning task according to the business processing request.
[0049] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps in the above method when executing the computer program.
[0050] A computer-readable storage medium stores a computer program, which implements the steps in the above method when executed by a processor.
[0051] The above-mentioned blacklist generation method, device, computer equipment and storage medium cluster the capital accounts through the capital transaction records, determine the risk category of the corresponding cluster, and then obtain the risk scoring model corresponding to the risk category, score the capital accounts through the risk scoring model, add the capital accounts with risk scores greater than the first threshold to the first-level nodes of the blacklist, add the associated accounts of the capital accounts of the first-level nodes to the second-level nodes of the blacklist, and finally obtain a blacklist divided into multiple levels according to the risk size of the capital accounts. Compared with the traditional method, the blacklist not only contains only the first-level nodes with higher risk scores, but also adds the associated accounts to the second-level nodes, fully considering the association relationship between risk accounts, thereby reducing the possibility of misjudgment and missed judgment, and effectively improving the accuracy of blacklist screening.
[0052] The above-mentioned blacklist screening method, device, computer equipment and storage medium extract pending business information and the requester's biometric information from the business processing request, and match the biometric information with the biometric information corresponding to the first-level node in the blacklist. When there is a matching first-level node, the matching degree of the pending business information and the transaction characteristics of the risk category corresponding to the first-level node is also calculated. Only when the matching degree is greater than the set value can it be determined that the business processing request hits the blacklist, and an early warning task is generated according to the business processing request. After the biometric information matches the biometric information corresponding to the first-level node in the blacklist, it is also necessary to match the pending business information with the corresponding transaction characteristics, so as to ensure that there will be no misjudgment or missed judgment during the blacklist screening process, and effectively improve the accuracy of blacklist screening. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] Figure 1 A schematic diagram of a process for generating a blacklist in an embodiment;
[0054] Figure 2 A schematic diagram of the fund transaction relationship in one embodiment;
[0055] Figure 3 A schematic diagram of a blacklist interface in one embodiment;
[0056] Figure 4 A diagram of an application environment of a blacklist screening method in one embodiment;
[0057] Figure 5 A schematic diagram of a process of a blacklist screening method in one embodiment;
[0058] Figure 6 is a structural block diagram of a blacklist generating device in one embodiment;
[0059] Figure 7 is a structural block diagram of a blacklist screening device in one embodiment;
[0060] Figure 8 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION
[0061] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0062] In one embodiment, Figure 1 As shown, a method for generating a blacklist is provided, comprising the following steps:
[0063] Step 102, obtaining fund transaction records.
[0064] Funds transaction records refer to the process records of customers using Internet finance to conduct specific business. Internet finance refers to a new financial business model in which traditional financial institutions and Internet companies use Internet technology and information and communication technology to achieve financing, payment, investment and information intermediary services. Funds transaction records include the inflow account of funds, the outflow account of funds, the amount of funds that have flowed or are flowing, etc. Funds transaction records can be the fund transaction records of some or all accounts for a certain or some specific business within a predetermined time period. Specific business may include transfer, recharge, withdrawal, etc., such as fund transaction records of all accounts for transfer within two months. The server can obtain fund transaction records of one or more accounts from the database, which stores fund transaction information of each account. The fund transaction records in the database can be updated in real time, and the transaction records are recorded in the database every time a fund transaction occurs; they can also be updated regularly at a certain frequency, and the fund transactions occurring during this period are recorded in the database at intervals. For example, it is updated every ten minutes.
[0065] Step 104, clustering the fund accounts according to the fund transaction records.
[0066] The server can use clustering algorithms to aggregate fund accounts into several clusters. Clustering algorithms are statistical analysis methods for studying classification problems, including partitioning methods, hierarchical methods, density algorithms, graph clustering methods, grid algorithms, model algorithms, etc. Fund accounts in a cluster have more similarities than fund accounts not in a cluster. Each cluster includes several fund accounts, and there may be overlaps between different clusters, that is, the same fund account may exist in one or more clusters.
[0067] In one of the embodiments, the fund accounts may be clustered according to the closeness of the fund account interactions, and multiple fund accounts that are closely interacting may be classified into the same cluster.
[0068] In one embodiment, the server may establish a fund transaction relationship diagram between fund accounts based on the fund transaction records. In the fund transaction relationship diagram, the fund accounts are points, and the fund transaction relationships are edges connecting the points. If a fund account exists in multiple fund transaction records, the same fund account is merged into the same point. The server clusters the fund accounts based on the fund transaction relationships in the fund transaction relationship diagram using a clustering algorithm. Figure 2As shown, ABCDEFGHIJK all represent capital accounts, and the lines between the circles represent the capital transaction relationship. The server uses a clustering algorithm to cluster the capital accounts into three clusters, namely cluster 1, cluster 2 and cluster 3. That is, ABCDE belongs to cluster 1, AHIL belongs to cluster 2, and FGJK belongs to cluster 3.
[0069] Step 106, obtaining transaction features of each cluster.
[0070] The server obtains the transaction features of each cluster, and each cluster has transaction features. The server can obtain transaction features from the fund transaction relationship between the fund accounts of the cluster, or from the fund account features. Specifically, the transaction features in the fund transaction relationship include transaction frequency features, transaction amount features, transaction number features, fund inflow and outflow features, etc. For example, the transaction feature can be that the fund transaction occurs at a fixed frequency every month or every quarter; the transaction feature can also be that the transaction amount is large and fixed each time, such as RMB 500,000; the transaction feature can also be that the funds are dispersed from multiple fund accounts into one fund account, and then the funds are transferred out of the fund account in a centralized manner. The transaction features from the fund account features include the concentration of the location of the fund account, the age stage of the fund account, etc. The server obtains the transaction features of each cluster from the fund transaction relationship or the fund account features, and the transaction features are the common similarities of the fund accounts or fund transaction relationships in the cluster.
[0071] Step 108: Determine the risk category of the corresponding cluster according to the transaction characteristics.
[0072] Among them, the risk category is the category of illegal and criminal behavior that may exist in the corresponding cluster, and the risk categories include money laundering category, pseudo-charity category, drug trafficking category, illegal fund-raising category, etc. A set of transaction features corresponding to each risk category is pre-defined. According to the corresponding relationship between transaction features and risk categories, the server can determine the risk category of the corresponding cluster based on the acquired transaction features. For example, the transaction features corresponding to the drug trafficking category include regular fund transactions between multiple fund accounts and the same fund account, that is, fund transactions are conducted at a certain frequency, and the amount of funds is stable and large or even huge. When the server obtains the transaction features of the cluster with a fixed frequency, a stable amount of funds, and a large amount, it determines the cluster as a drug trafficking category, such as Figure 2 Clustering in 1. When the server obtains the transaction feature of the cluster that funds are traded multiple times between fixed fund accounts and finally concentrated in one fund account, the cluster is determined as a money laundering category, such as Figure 2 Cluster 2 in .
[0073] Step 110: Obtain a risk scoring model corresponding to the risk category.
[0074] Each risk category has a corresponding risk scoring model, for example, the drug trafficking category has a corresponding drug trafficking risk scoring model, and the illegal fund-raising category has a corresponding illegal fund-raising risk scoring model. The server can obtain the corresponding risk scoring model according to the determined risk category.
[0075] Step 112, scoring the fund accounts in the corresponding cluster using the risk scoring model to obtain a risk score.
[0076] The server uses the risk scoring model to score all the fund accounts in the corresponding cluster one by one to obtain the risk score. Specifically, the server obtains the standard transaction features of the corresponding risk category preset in the risk scoring model, compares the standard transaction features with the transaction features of the fund account one by one, and scores them according to the similarity. The more similar the fund account's transaction features are to the standard transaction features, the higher the risk score is, indicating that it is more likely to belong to the risk category; the more different the fund account's transaction features are from the standard transaction features, the lower the risk score is, indicating that it is less likely to belong to the risk category.
[0077] Step 114, adding the fund accounts with risk scores greater than the first threshold to the first-level nodes of the blacklist.
[0078] The server adds the capital account with a risk score greater than the first threshold to the first-level node of the blacklist. The blacklist includes the first-level node, the second-level node, the third-level node, etc. The first threshold is the preset first risk score. When the risk score of the capital account is greater than the first threshold, the server adds the capital account to the first-level node of the blacklist. Specifically, the server adds the basic information of the account subject corresponding to the capital account to the first-level node of the blacklist. When the account subject corresponding to the capital account is a natural person, the basic information includes the basic identity information of the natural person, such as name, gender, age, nationality, ID number, place of household registration, home address, etc.; when the account subject corresponding to the capital account is a legal person, the basic information includes the basic information of the legal person and the basic identity information of the legal representative, such as the name of the legal person, place of registration, business number and the name, gender, age, nationality, ID number, place of household registration, home address, etc. of the legal representative.
[0079] Step 116, obtaining the associated accounts of the capital account of the first-level node from the cluster, and adding the associated accounts to the second-level nodes of the blacklist.
[0080] The server obtains the accounts with fund transactions from the clusters corresponding to the fund accounts of the first-level nodes as associated accounts, and adds the associated accounts to the second-level nodes of the blacklist. Specifically, the server also adds the basic information of the account principal corresponding to the associated account to the second-level nodes of the blacklist.
[0081] like Figure 2 , Figure 3 As shown, if the risk score of the capital account of A and I is greater than the first threshold, A and I are added to the first-level node of the blacklist. The server then obtains the capital accounts of the first-level node from clusters 1 and 2 to which A and I belong, that is, the associated accounts of A and I. The associated accounts of A include BCDEI, and the associated accounts of I include AHL. The server adds the basic information of the account subject corresponding to ABCDEHIL to the second-level node of the blacklist.
[0082] The above blacklist generation method clusters the capital accounts through the capital transaction records, determines the risk category of the corresponding cluster, and then obtains the risk scoring model corresponding to the risk category, scores the capital accounts through the risk scoring model, adds the capital accounts with risk scores greater than the first threshold to the first-level nodes of the blacklist, and adds the associated accounts of the capital accounts of the first-level nodes to the second-level nodes of the blacklist, and finally obtains a blacklist divided into multiple levels according to the risk size of the capital accounts. Compared with the traditional method, the blacklist not only contains only the first-level nodes with higher risk scores, but also adds the associated accounts to the second-level nodes, fully considering the association relationship between risk accounts, thereby reducing the possibility of misjudgment and missed judgment, and effectively improving the accuracy of blacklist screening.
[0083] In one embodiment, the blacklist generation method further includes the following steps:
[0084] Obtain the risk score of the capital account of the second-level node; compare the risk score with the second threshold, and determine the target second-level node whose risk score is greater than the second threshold; obtain the associated account of the capital account corresponding to the target second-level node from the cluster, and add the associated account to the third-level node of the blacklist, where the third-level node is the third-level node corresponding to the target second-level node.
[0085] The server obtains the risk score of each fund account in the second-level node of the blacklist, and compares the risk score with the second threshold one by one. Among them, the second threshold is a pre-set second risk score, and the second threshold is less than the first threshold. The server selects the fund account set with a risk score greater than the second threshold from the second-level node of the blacklist as the target second-level node, and the fund account in the target second-level node corresponds to the corresponding cluster. The server obtains the fund account associated with the fund account corresponding to the target second-level node from the cluster as the associated account, and adds the associated account to the third-level node of the blacklist. It can be understood that the risk score corresponding to the associated account in the third-level node is greater than the second threshold and less than the first threshold. If the risk score corresponding to the fund account in the second-level node is greater than the second threshold and greater than the first threshold, the fund account belongs to the first-level node, so the associated account of the fund account is no longer obtained from the cluster and added to the third-level node of the blacklist. Specifically, the server adds the associated account to the third-level node by adding the basic information of the account subject corresponding to the associated account to the third-level node of the blacklist.
[0086] In this embodiment, the risk score of the fund account of the second-level node of the blacklist is compared with the second threshold value, the target second-level node is determined, and then the associated account of the fund account corresponding to the target second-level node is added to the third-level node of the blacklist. The entire blacklist is divided into a multi-level structure of first-level nodes, second-level nodes, and third-level nodes according to the risk score and the association relationship. The possibility of the fund account in the first-level node of the blacklist being at risk is higher, and the possibility of the fund account in the third-level node being at risk is relatively small. In addition to taking the risk score as the standard, the method for generating the blacklist also takes into account the associated accounts associated with the fund account with a high risk score. With the blacklist generated in this way, the risk level of the fund account is clear and the radiation range is wide, which can effectively reduce the possibility of misjudgment and missed judgment, thereby improving the accuracy of blacklist screening.
[0087] In one embodiment, the blacklist generation method further includes the following steps:
[0088] Collect the biometric information of the account holders of all fund accounts in the blacklist; add the biometric information to the corresponding node of the blacklist.
[0089] The server collects the biometric information of the account subjects of all the capital accounts in the blacklist. The account subjects of the capital accounts include natural persons and legal persons. When the account subject of the capital account is a natural person, the server collects the biometric information of the natural person. The biometric information of the natural person includes facial feature information, voiceprint information, fingerprint information, pulse information, iris information, etc. When the account subject of the capital account is a legal person, the server collects the biometric information of the legal representative of the legal person. The biometric information of the legal representative includes facial feature information, voiceprint information, fingerprint information, pulse information, iris information, etc. The server adds the collected biometric information of the account subject to the node of the blacklist, so that the biometric information corresponds to the capital account in the blacklist node.
[0090] In this embodiment, the blacklist includes not only the basic information of the account subject of the capital account, but also the biometric information of the account subject. Compared with the traditional blacklist that only includes basic information, the information of the capital account in the blacklist generated in this embodiment is richer, and the biometric information is unique to each account subject. The capital account in the blacklist corresponds to the biometric information, which effectively prevents the situation where the matching result is inaccurate only through basic information, thereby reducing the possibility of "false positive" and "false negative" misjudgments and improving the accuracy of blacklist screening.
[0091] The blacklist screening method provided in this application can be applied to Figure 4 In the application environment shown. Among them, the terminal 402 communicates with the server 404 through the network. The terminal 402 can send a business processing request to the server 404, and the server 404 extracts the pending business information and the biometric information of the requester from the business processing request, and determines whether the business processing request hits the blacklist through the matching result of the biometric information of the requester and the biometric information in the blacklist, and the matching degree of the pending business information and the transaction characteristics of the risk category. Among them, the terminal 402 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers and portable wearable devices, and the server 404 can be implemented by an independent server or a server cluster composed of multiple servers.
[0092] In one embodiment, Figure 5 As shown, a blacklist screening method is provided, and the method is applied to Figure 4 The server in the example is used to illustrate the following steps:
[0093] Step 502: Receive a service processing request.
[0094] Step 504: extract the pending service information and the requester's biometric information from the service processing request.
[0095] The server receives a business processing request uploaded by the terminal. The business processing request is a request uploaded by the user through the terminal to handle Internet financial services. For example, a transfer request, a withdrawal request, a recharge request, etc. The business processing request includes pending business information and the biometric information of the requester. Among them, the pending business information may include the pending business type, the pending business amount, the transaction object information, etc.; the biometric information of the requester may include one or more of the facial image, fingerprint information, iris information, etc. uploaded by the requester. When the requester is a legal person, the biometric information of the legal representative is uploaded. The server extracts the pending business information and the biometric information of the requester from the business processing request.
[0096] Step 506 , feature matching is performed between the biometric information and the biometric information corresponding to the first-level node in the blacklist. If there is a matching first-level node, the risk category corresponding to the first-level node is obtained.
[0097] The server performs feature matching on the biometric information extracted from the business processing request with the biometric information corresponding to the first-level node in the blacklist one by one. The server performs feature matching on each type of information in the biometric information with the biometric information corresponding to the first-level node in the blacklist one by one. When at least one feature information in the biometric information is matched successfully, it is determined that there is a fund account in the first-level node in the blacklist that matches the biometric information of the requester. The server obtains the risk category corresponding to the fund account of the matched first-level node, which is the risk category of the cluster to which the fund account belongs.
[0098] In one embodiment, the server matches the biometric information item by item, and obtains the final matching result after weighting each matching result. For example, the biometric information of the requester carried in the business processing request includes facial image, fingerprint and voiceprint. The server matches the facial image, fingerprint and voiceprint with the facial image, fingerprint and voiceprint in the first-level node in the blacklist, respectively, to obtain the facial image matching result, fingerprint matching result and voiceprint matching result. The server performs weighted calculation on the facial image matching result, fingerprint matching result and voiceprint matching result according to the preset weight of each biometric category to obtain the final matching result. According to the final matching result, it is judged whether there is a fund account matching with the requester. In the traditional method, the screening of the blacklist only screens the basic identity information such as the name and ID number of the requester, and the offender can easily avoid the screening of the blacklist by creating a false identity. In this embodiment, the matching is carried out through the biometric information of multiple dimensions, which effectively avoids the "false negative" misjudgment caused by identity fraud, and effectively improves the accuracy of the blacklist screening.
[0099] Step 508, calculating the matching degree between the pending business information and the transaction characteristics of the risk category. If the matching degree is greater than a set value, it is determined that the business processing request hits the blacklist.
[0100] The server obtains the transaction features corresponding to the risk categories, matches the pending business information in the business processing request with the transaction features, and calculates the matching degree between them. If the matching degree is greater than the set value, the business processing request is determined to have hit the blacklist. If the capital account of the matching first-level node belongs to multiple clusters, the transaction features corresponding to the risk categories of each cluster are obtained respectively, and the pending business information is matched with the transaction features one by one. If any matching degree is greater than the set value, the business processing request is determined to have hit the blacklist. Figure 2 As shown, if A is the capital account of the first-level node that matches the biometric information of the requester, and A belongs to cluster 1 and cluster 2, the server obtains the risk categories of cluster 1 and cluster 2 respectively. If the risk category of cluster 1 is drug trafficking, and the risk category of cluster 2 is illegal fund-raising, the server obtains the transaction features of drug trafficking and illegal fund-raising respectively, and matches the pending business information in the business processing request with the transaction features of drug trafficking and illegal fund-raising respectively. If the matching degree of any one of the drug trafficking category or illegal fund-raising category is greater than the set value, it is determined that the business processing request hits the blacklist.
[0101] Step 510: Generate an early warning task according to the business processing request.
[0102] The server generates an early warning task based on the business processing request that hits the blacklist, and the early warning task includes the risk category that the business processing request may have, the first-level node of the blacklist that hits, etc. The server can send the generated early warning task to the business processing terminal for processing to remind the processing personnel that the business processing request has hit the blacklist.
[0103] The above-mentioned blacklist screening method extracts pending business information and the biometric information of the requester from the business processing request, and matches the biometric information with the biometric information corresponding to the first-level node in the blacklist. When there is a matching first-level node, the matching degree of the pending business information and the transaction characteristics of the risk category corresponding to the first-level node is also calculated. Only when the matching degree is greater than the set value can it be determined that the business processing request hits the blacklist, and an early warning task is generated according to the business processing request. After the biometric information matches the biometric information corresponding to the first-level node in the blacklist, it is also necessary to match the pending business information with the corresponding transaction characteristics, so as to ensure that there will be no misjudgment or missed judgment during the blacklist screening process, which effectively improves the accuracy of blacklist screening.
[0104] In one embodiment, after the step 506 of matching the biometric information with the biometric information corresponding to the first-level node in the blacklist, the method further includes:
[0105] If there is no matching first-level node, the biometric information is feature matched with the biometric information corresponding to the second-level node in the blacklist; if there is a matching second-level node, the biometric information of the first-level node corresponding to the matching second-level node is obtained; the biometric information of the transaction object is extracted from the business processing request; the biometric information of the transaction object is feature matched with the biometric information of the first-level node; if there is a matching first-level node, it is determined that the business processing request hits the blacklist.
[0106] If the server does not find a first-level node that matches the biometric information of the requester, it will feature match the biometric information with the biometric information corresponding to the second-level node in the blacklist. The specific feature matching method is similar to the above-mentioned feature matching, so it will not be repeated here. If there is a capital account of a second-level node that matches the biometric information of the requester, the biometric information of the first-level node corresponding to the capital account of the second-level node is obtained. The second-level node is composed of the associated accounts of the capital account of the first-level node, so each capital account of the second-level node has a corresponding first-level node. The server extracts the biometric information of the transaction object from the business processing request, and the business processing request carries at least one biometric information of the transaction object. The server features match the biometric information of the transaction object with the biometric information of the first-level node. If there is a matching first-level node, it is determined that the business processing request hits the blacklist.
[0107] In this embodiment, the second-level node is a low-risk node, that is, a node associated with the first-level node, and the second-level node is a low-risk node relative to the first-level node. In the traditional blacklist screening method, when the requester does not belong to the blacklist, the blacklist is not hit. The traditional blacklist low-risk personnel are not included in the blacklist. Therefore, it is easy to miss the "false negative" misjudgment of low-risk personnel. In this embodiment, the capital account associated with the high-risk first-level node is also added to the blacklist. The requester does not belong to the high-risk user, but the requester is an associated user of the high-risk user. The traditional blacklist does not hit this type of business request. In this embodiment, if the transaction object of the requester is a high-risk user who has conducted risky transactions with the requester, the business request hits, realizing the financial transactions between low-risk personnel and high-risk personnel. The multi-level blacklist increases the breadth of risk request hits and effectively improves the accuracy of blacklist screening.
[0108] In one embodiment, the server performs feature matching between the biometric information of the requester and the biometric information corresponding to the second-level node in the blacklist. If there is a matching fund account of the second-level node, the server can also obtain the risk category of the cluster to which the fund account of the second-level node belongs, and calculate the degree of match between the pending business information and the transaction characteristics of the risk category. If the degree of match is greater than a set value, it is determined that the business processing request hits the blacklist.
[0109] In this embodiment, after the server finds the matching fund account of the second-level node, it determines whether the pending business information may belong to the risk category corresponding to the fund account of the second-level node. That is, when the server matches a low-risk node, because the risk of the node is relatively low, it continues to determine whether the pending business information is a corresponding illegal or criminal behavior. If so, it is determined that the business processing request hits the blacklist; if not, it is determined that the business processing request does not hit the blacklist. In this embodiment, the crime possibility of the low-risk node, that is, the second-level node is relatively small. After the biometric information of the applicant hits the second-level node, it is also necessary to determine whether the pending business belongs to an illegal or criminal behavior, which effectively avoids the occurrence of "false positive" misjudgments and improves the accuracy of blacklist screening.
[0110] In one embodiment, if the server does not find a first-level node fund account that matches the biometric information of the applicant in the first-level node of the blacklist, the server performs feature matching on the biometric information of the applicant with the biometric information corresponding to the second-level node in the blacklist, and obtains the biometric information of the third-level node corresponding to the matching second-level node; performs feature matching on the biometric information of the transaction object with the biometric information corresponding to the third-level node, and if there is a matching third-level node, obtains the risk category corresponding to the third-level node; calculates the matching degree of the transaction characteristics of the pending business information and the risk category, and if the matching degree is greater than the set value, determines that the business processing request hits the blacklist.
[0111] The server obtains the biometric information of the third-level node, which is the third-level node corresponding to the second-level node that matches the biometric information of the requester. The server performs feature matching between the biometric information of the transaction object and the biometric information corresponding to the third-level node. The capital account of the third-level node is the associated account of the capital account of the second-level node. The biometric information of the transaction object is matched with the third-level node to determine whether the capital transaction is a transaction between the second-level node and the third-level node. If so, the risk category corresponding to the matched third-level node is obtained, and the matching degree between the pending business information and the transaction characteristics of the risk category is calculated. If the matching degree is greater than the set value, it is determined that the business processing request hits the blacklist.
[0112] In this embodiment, because the crime risk of the second-level node and the third-level node is relatively small, after the requester and the transaction object are matched with the second-level node and the third-level node respectively, the matching degree of the pending business information with the transaction characteristics of the risk category corresponding to the third-level node is calculated. If the matching degree is greater than the set value, it is determined that the business processing request hits the blacklist. In this embodiment, not only is it determined whether the requester and the transaction object match the second-level node and the third-level node, but it is also determined whether the pending business belongs to the risk category. Only when all are satisfied, it is determined that the business processing request hits the blacklist. At the same time, compared with the traditional method, this embodiment screens the second-level nodes and the third-level nodes with very small crime risks. It greatly reduces the "false negative" and "false positive" misjudgments and missed judgments, and improves the accuracy of blacklist screening.
[0113] It should be understood that although Figure 1 , Figure 5 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 , Figure 5 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.
[0114] In one embodiment, Figure 6 As shown, a blacklist generation device is provided, including: an acquisition module 602, a clustering module 604, a scoring module 606 and a blacklist generation module 608, wherein:
[0115] The acquisition module 602 is used to acquire fund transaction records.
[0116] The clustering module 604 is used to cluster the fund accounts according to the fund transaction records.
[0117] The acquisition module 602 is also used to acquire transaction features of each cluster; determine the risk category of the corresponding cluster according to the transaction features; and acquire the risk scoring model corresponding to the risk category.
[0118] The scoring module 606 is used to score the fund accounts in the corresponding cluster using a risk scoring model to obtain a risk score.
[0119] The blacklist generation module 608 is used to add the capital accounts with risk scores greater than the first threshold to the first-level nodes of the blacklist; obtain the associated accounts of the capital accounts of the first-level nodes from the cluster, and add the associated accounts to the second-level nodes of the blacklist.
[0120] In one embodiment, the acquisition module 602 is further used to acquire the risk score of the fund account of the second-level node; compare the risk score with the second threshold, and determine the target second-level node whose risk score is greater than the second threshold.
[0121] The blacklist generation module 608 is also used to obtain the associated account of the fund account corresponding to the target second-level node from the cluster, and add the associated account to the third-level node of the blacklist, where the third-level node is the third-level node corresponding to the target second-level node.
[0122] In one embodiment, the blacklist generation device further includes a biometric feature adding module, wherein the biometric feature adding module is used to collect the biometric feature information of the account holders of all the fund accounts in the blacklist, and add the biometric feature information to the corresponding node of the blacklist.
[0123] In one embodiment, Figure 7 As shown, a blacklist screening device is provided, including: a request receiving module 702, an information extraction module 704, a feature matching module 706, a determination module 708 and a task generation module 710, wherein:
[0124] The request receiving module 702 is used to receive a service processing request.
[0125] The information extraction module 704 is used to extract pending business information and the biometric information of the requester from the business processing request.
[0126] The feature matching module 706 is used to perform feature matching between the biometric information and the biometric information corresponding to the first-level node in the blacklist, and if there is a matching first-level node, obtain the risk category corresponding to the first-level node.
[0127] The determination module 708 is used to calculate the matching degree between the pending business information and the transaction characteristics of the risk category. If the matching degree is greater than a set value, it is determined that the business processing request hits the blacklist.
[0128] The task generation module 710 is used to generate an early warning task according to a business processing request.
[0129] In one embodiment, the feature matching module 706 is also used to perform feature matching between the biometric information and the biometric information corresponding to the second-level node in the blacklist if there is no matching first-level node; if there is a matching second-level node, obtain the biometric information of the first-level node corresponding to the matching second-level node; extract the biometric information of the transaction object from the business processing request; and perform feature matching between the biometric information of the transaction object and the biometric information of the first-level node.
[0130] The above-mentioned determination module 708 is also used to determine whether the business processing request hits the blacklist if there is a matching first-level node.
[0131] In one embodiment, the information extraction module 704 is further used to obtain biometric information of a third-level node corresponding to the matched second-level node;
[0132] The feature matching module 706 is also used to perform feature matching between the biometric information of the transaction object and the biometric information corresponding to the third-level node. If there is a matching third-level node, the risk category corresponding to the third-level node is obtained.
[0133] The above-mentioned determination module 708 is also used to calculate the matching degree between the pending business information and the transaction characteristics of the risk category. If the matching degree is greater than a set value, it is determined that the business processing request hits the blacklist.
[0134] For the specific definition of the blacklist generation and screening device, please refer to the definition of the blacklist generation and screening method above, which will not be repeated here. Each module in the above-mentioned blacklist generation and screening device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
[0135] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store blacklist generation and screening data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a blacklist generation and screening method is implemented.
[0136] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0137] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiment are implemented.
[0138] It should be noted that the user information (including but not limited to user device information, user personal information such as biometric information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.
[0139] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0140] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0141] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.
Claims
1. A method for generating a blacklist, characterized in that: The method comprises: Obtain fund transaction records; Clustering the fund accounts according to the fund transaction records; Obtaining transaction features of each of the clusters; Determining the risk category of the corresponding cluster according to the transaction characteristics; Obtaining a risk scoring model corresponding to the risk category; Scoring the fund accounts in the corresponding cluster using the risk scoring model to obtain a risk score; Adding the fund account having the risk score greater than a first threshold to the first-level node of the blacklist; Obtaining the associated account of the fund account of the first-level node from the cluster, and adding the associated account to the second-level node of the blacklist; the blacklist is used to perform feature matching between the biometric information corresponding to the first-level node in the blacklist and the biometric information of the requester, and if there is a matching first-level node, obtaining the risk category corresponding to the first-level node; calculating the matching degree of the transaction characteristics of the pending business information and the risk category corresponding to the first-level node, and if the matching degree is greater than a set value, determining that the business processing request hits the blacklist; generating an early warning task according to the business processing request; If there is no matching first-level node, the biometric information of the requester is feature matched with the biometric information corresponding to the second-level node in the blacklist; if there is a matching second-level node, the biometric information of the first-level node corresponding to the matching second-level node is obtained; the biometric information of the transaction object is extracted from the business processing request; the biometric information of the transaction object is feature matched with the biometric information of the first-level node; if there is a matching first-level node, it is determined that the business processing request hits the blacklist; the biometric information of the third-level node corresponding to the matching second-level node is obtained; the biometric information of the transaction object is feature matched with the biometric information corresponding to the third-level node, and if there is a matching third-level node, the risk category corresponding to the third-level node is obtained; the matching degree of the transaction characteristics of the pending business information and the risk category corresponding to the third-level node is calculated, and if the matching degree is greater than a set value, it is determined that the business processing request hits the blacklist; an early warning task is generated according to the business processing request, and the pending business information and the biometric information of the requester are extracted from the business processing request based on receiving the business processing request.
2. The method according to claim 1, characterized in that The method further comprises: Obtaining the risk score of the fund account of the second-level node; Compare the risk score of the fund account of the second-level node with a second threshold, and determine a target second-level node whose risk score is greater than the second threshold; Obtain an associated account of the capital account corresponding to the target second-level node from the cluster, and add the associated account to a third-level node of the blacklist, where the third-level node is a third-level node corresponding to the target second-level node.
3. The method according to claim 1, characterized in that: The method further comprises: Collecting biometric information of account holders of all fund accounts in the blacklist; The biometric information of the account holders of all the fund accounts in the blacklist is added to the corresponding nodes of the blacklist.
4. A blacklist screening method, characterized in that: The method comprises: Receive business processing requests; Extracting pending business information and biometric information of the requester from the business processing request; Perform feature matching on the biometric information and the biometric information corresponding to the first-level node in the blacklist. If there is a matching first-level node, obtain the risk category corresponding to the first-level node; calculate the matching degree of the transaction characteristics of the pending business information and the risk category corresponding to the first-level node. If the matching degree is greater than a set value, determine that the business processing request hits the blacklist; generate an early warning task according to the business processing request; the blacklist is generated by obtaining fund transaction records; clustering fund accounts according to the fund transaction records; obtaining transaction characteristics of each cluster; determining the risk category of the corresponding cluster according to the transaction characteristics; obtaining a risk scoring model corresponding to the risk category of the cluster; scoring the fund accounts in the corresponding cluster by the risk scoring model to obtain a risk score; adding the fund account with a risk score greater than a first threshold to the first-level node of the blacklist; obtaining an associated account of the fund account of the first-level node from the cluster, and adding the associated account to the second-level node of the blacklist; If there is no matching first-level node, feature matching is performed between the biometric information and the biometric information corresponding to the second-level node in the blacklist; If there is a matching second-level node, obtaining the biometric information of the first-level node corresponding to the matching second-level node; Extracting biometric information of the transaction object from the business processing request; Perform feature matching on the biometric information of the transaction object and the biometric information of the first-level node; If there is a matching first-level node, determining that the service processing request hits the blacklist; Acquire biometric information of a third-level node corresponding to the matched second-level node; Perform feature matching on the biometric information of the transaction object and the biometric information corresponding to the third-level node, and if there is a matching third-level node, obtain the risk category corresponding to the third-level node; The matching degree between the pending business information and the transaction characteristics of the risk category corresponding to the third-level node is calculated. If the matching degree is greater than a set value, it is determined that the business processing request hits the blacklist; and an early warning task is generated according to the business processing request.
5. The method according to claim 4, characterized in that The second-level nodes are low-risk nodes relative to the first-level nodes.
6. The method according to claim 4, characterized in that The capital account of the third-level node is an associated account of the capital account of the second-level node.
7. A blacklist generation device, characterized in that: The device comprises: Acquisition module, used to obtain fund transaction records; A clustering module, used for clustering fund accounts according to the fund transaction records; The acquisition module is also used to acquire transaction features of each cluster; determine the risk category of the corresponding cluster according to the transaction features; and acquire the risk scoring model corresponding to the risk category; A scoring module, used to score the fund accounts in the corresponding cluster using the risk scoring model to obtain a risk score; A blacklist generation module, used for adding the fund account whose risk score is greater than a first threshold to the first-level node of the blacklist; obtaining the associated account of the fund account of the first-level node from the cluster, and adding the associated account to the second-level node of the blacklist; the blacklist is used for feature matching between the biometric information corresponding to the first-level node in the blacklist and the biometric information of the requester, and if there is a matching first-level node, obtaining the risk category corresponding to the first-level node; calculating the matching degree between the pending business information and the transaction characteristics of the risk category corresponding to the first-level node, and if the matching degree is greater than a set value, determining that the business processing request hits the blacklist; generating an early warning task according to the business processing request; If there is no matching first-level node, feature matching is performed on the biometric information of the applicant with the biometric information corresponding to the second-level node in the blacklist; if there is a matching second-level node, the biometric information of the first-level node corresponding to the matching second-level node is obtained; the biometric information of the transaction object is extracted from the business processing request; the biometric information of the transaction object is feature matched with the biometric information of the first-level node; if there is a matching first-level node, it is determined that the business processing request hits the blacklist; the biometric information of the third-level node corresponding to the matching second-level node is obtained; the biometric information of the transaction object is matched; Perform feature matching with the biometric information corresponding to the third-level node, and if a matching third-level node exists, obtain the risk category corresponding to the third-level node; calculate the matching degree of the transaction features of the pending business information and the risk category corresponding to the third-level node, and if the matching degree is greater than a set value, determine that the business processing request hits the blacklist; calculate the matching degree of the transaction features of the pending business information and the risk category, and if the matching degree is greater than a set value, determine that the business processing request hits the blacklist; generate an early warning task according to the business processing request, the pending business information and the requester's biometric information are extracted from the business processing request based on receiving the business processing request.
8. A blacklist screening device, characterized in that: The device comprises: A request receiving module, used to receive a business processing request; An information extraction module, used to extract pending business information and biometric information of the requester from the business processing request; A feature matching module is used to perform feature matching between the biometric feature information and the biometric feature information corresponding to the first-level node in the blacklist, and if there is a matching first-level node, obtain the risk category corresponding to the first-level node; the blacklist is generated by obtaining fund transaction records; clustering fund accounts according to the fund transaction records; obtaining transaction features of each cluster; determining the risk category of the corresponding cluster according to the transaction features; obtaining a risk scoring model corresponding to the risk category of the cluster; scoring the fund accounts in the corresponding cluster by the risk scoring model to obtain a risk score; adding the fund accounts with a risk score greater than a first threshold to the first-level node of the blacklist; obtaining an associated account of the fund account of the first-level node from the cluster, and adding the associated account to the second-level node of the blacklist; If there is no matching first-level node, feature matching is performed on the biometric information with the biometric information corresponding to the second-level node in the blacklist; if there is a matching second-level node, the biometric information of the first-level node corresponding to the matching second-level node is obtained; the biometric information of the transaction object is extracted from the business processing request; the biometric information of the transaction object is feature matched with the biometric information of the first-level node; if there is a matching first-level node, it is determined that the business processing request hits the blacklist; the biometric information of the third-level node corresponding to the matching second-level node is obtained; the biometric information of the transaction object is feature matched with the biometric information corresponding to the third-level node, and if there is a matching third-level node, the risk category corresponding to the third-level node is obtained; A determination module, configured to calculate the matching degree between the pending business information and the transaction characteristics of the risk category corresponding to the first-level node, and if the matching degree is greater than a set value, determine that the business processing request hits the blacklist; calculate the matching degree between the pending business information and the transaction characteristics of the risk category corresponding to the third-level node, and if the matching degree is greater than a set value, determine that the business processing request hits the blacklist; The task generation module is used to generate an early warning task according to the business processing request.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 or 4 to 6 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 or 4 to 6 are implemented.
Citation Information
Patent Citations
Method and device for identifying risks and system for preventing and controlling same
CN106355405A
Blacklist data generating method and device
CN106776973A
Habitual offender fraud detection method and system in finance credit audit business
CN107563758A
Method and device for establishing phone number blacklist based on a communication network
CN108055661A
Data processing method, apparatus and device for insurance fraud identification, and server
CN108734479A
