A method and device for monitoring capital flow based on dual-graph fusion calculation

Abstract

The present invention belongs to the field of big data technology. The present invention provides a method and device for monitoring fund flow based on dual-graph fusion calculation. The method for monitoring fund flow based on dual-graph fusion calculation includes: generating a customer association relationship map of the target customer based on the target customer ID and the customer association relationship of the target customer; generating a fund flow time series map according to the fund flow direction and fund relationship of the target customer; monitoring the fund flow of the target customer according to the customer association relationship map and the fund flow time series map. The present invention combines the dynamic relationship reflecting the transaction timing and flow direction with the static relationship reflecting the customer association relationship through the dual-graph fusion calculation idea, breaking through the limitation of traditional fund flow relationship monitoring that only focuses on fund transactions, and realizing the exploration of implicit fund relationships.

Description

Technical Field

[0001] The present application belongs to the field of big data technology, and specifically relates to a method and device for monitoring capital flow based on dual-graph fusion calculation. Background Art

[0002] In the financial field, monitoring and tracking of capital flows is an important means for financial institutions to analyze whether the source and use of customer funds are legal and compliant. It has very important application value in credit, anti-fraud, anti-money laundering and other business fields. However, based on existing technologies, financial institutions are limited by the scope of available data and can only analyze a fragment of the entire chain of capital flows. They cannot conduct effective analysis and gain insight into the full picture of capital chain breaks caused by cash or inter-bank transactions. More and more illegal customers are deliberately evading bank monitoring in increasingly complex transaction methods, so there is an urgent need for a set of monitoring methods and ideas that break through data limitations, so as to restore more hidden capital flow relationships for business personnel and reveal more business risks. On the other hand, in order to complicate the capital chain, illegal customers generally rely on trusted parties. This provides new ideas for business personnel to expand capital relationships, build customer relationship circles, and expand capital chains. Summary of the invention

[0003] The present invention can be used in the technical field of big data application in finance, and can also be used in any field other than the financial field. The application field of the fund flow monitoring method and device based on dual-spectrum fusion calculation disclosed in the present invention is not limited. The present invention combines the dynamic relationship reflecting the transaction sequence and flow direction with the static relationship reflecting the customer association relationship through the dual-spectrum fusion calculation idea, breaking through the limitation of traditional fund flow relationship monitoring that only focuses on fund transactions, and realizing the exploration of implicit fund relationships.

[0004] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0005] In a first aspect, the present invention provides a method for monitoring capital flow based on dual-graph fusion calculation, comprising:

[0006] Generate a customer association relationship map of the target customer based on the target customer ID and the customer association relationship of the target customer;

[0007] Generate a funds flow time series graph based on the funds flow and funds relationship of the target customer;

[0008] The fund flow of the target customer is monitored according to the customer association relationship map and the fund flow time sequence map.

[0009] In one embodiment, the step of generating a customer association relationship map of the target customer based on the target customer ID and the customer association relationship of the target customer includes:

[0010] Determine the target customer and the target customer ID based on the business scenario;

[0011] Using a graph database to perform connectivity detection and loop mining on the target customer and its related customers to determine the customer association relationship;

[0012] The customer association relationship map is generated according to the target customer ID, the related customer ID and the customer association relationship.

[0013] In one embodiment, generating a capital flow time series graph according to the capital flow and capital relationship of the target customer includes:

[0014] Determine the data scope of the target customer based on the monitoring scenario;

[0015] Within the data range, classifying the nodes in the capital flow time series graph according to the counterparty category of the target customer;

[0016] Marking the edges in the capital flow time series graph formed by normal capital flow;

[0017] The fund flow time series graph is generated according to the classified nodes and the marked edges.

[0018] In one embodiment, monitoring the target customer's fund flow according to the customer association relationship map and the fund flow time sequence map includes:

[0019] Determine the initial edge of the target customer in the capital flow time series graph, and perform the following iterative operations:

[0020] Determining whether the transaction counterparty is an entity in the capital flow time series graph according to the first multi-degree capital relationship;

[0021] If yes, searching the customer relationship of the counterparty in the customer relationship graph;

[0022] According to the customer association relationship of the transaction counterparty and the second multi-degree capital relationship, it is determined in the capital flow time series graph whether the transaction counterparty corresponding to the transaction counterparty is an entity, until all customers in the customer association relationship graph are calculated.

[0023] In a second aspect, the present invention provides a fund flow monitoring device based on dual-graph fusion calculation, the device comprising:

[0024] A customer relationship map generation module, used to generate a customer relationship map of the target customer based on the target customer ID and the customer relationship of the target customer;

[0025] A fund flow graph generation module is used to generate a fund flow time series graph according to the fund flow and fund relationship of the target customer;

[0026] The fund flow monitoring module is used to monitor the fund flow of the target customer based on the customer association relationship map and the fund flow time sequence map.

[0027] In one embodiment, the customer relationship map generation module includes:

[0028] A customer determination unit, configured to determine the target customer and the target customer ID based on a business scenario;

[0029] An association relationship determination unit, configured to perform connectivity detection and loop mining on the target customer and its related customers using a graph database to determine the customer association relationship;

[0030] A customer relationship map generating unit is used to generate the customer relationship map according to the target customer ID, the related customer ID and the customer relationship.

[0031] In one embodiment, the fund flow graph generation module includes:

[0032] A data range determination unit, used to determine the data range of the target customer based on a monitoring scenario;

[0033] A node classification unit, used for classifying nodes in the capital flow time series graph according to the counterparty category of the target customer within the data range;

[0034] An edge identification unit, used for identifying edges in the capital flow time series graph formed by normal capital flow;

[0035] The capital flow graph generation unit is used to generate the capital flow time series graph according to the classified nodes and the marked edges.

[0036] In one embodiment, the fund flow monitoring module includes:

[0037] The initial edge determination unit is used to determine the initial edge of the target customer in the capital flow time series graph and perform the following iterative operations:

[0038] A transaction counterparty determination unit, configured to determine whether the transaction counterparty is an entity in the capital flow time series graph according to the first multi-degree capital relationship;

[0039] an association relationship search unit, configured to search for customer association relationships of the transaction counterparty in the customer association relationship map;

[0040] The counterparty judgment subunit is used to judge whether the counterparty corresponding to the counterparty in the capital flow time series map is an entity according to the customer relationship of the counterparty and the second multi-degree capital relationship, until all customers in the customer relationship map are calculated.

[0041] In a third aspect, the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of a method for monitoring capital flow based on dual-graph fusion calculation are implemented.

[0042] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for monitoring capital flow based on dual-graph fusion calculation.

[0043] From the above description, it can be seen that the embodiment of the present invention provides a method and device for monitoring capital flow based on dual-graph fusion calculation. First, a customer association relationship graph of the target customer is generated based on the target customer ID and the customer association relationship of the target customer; then, a capital flow time series graph is generated according to the capital flow direction and capital relationship of the target customer; finally, the capital flow of the target customer is monitored according to the customer association relationship graph and the capital flow time series graph. The present invention, through the dual-spectrum fusion calculation idea, combines the dynamic relationship that reflects the transaction timing and flow direction with the static relationship that reflects the customer association relationship to form a calculation mode, breaking through the limitation of traditional capital flow relationship monitoring that only focuses on capital transactions, and realizes the exploration of implicit capital relationships. The beneficial effects of the present invention are as follows:

[0044] First, it breaks through the limitation of relying solely on transaction data to monitor cash flows. After integrating customer relationships, the search space is further expanded, the capital chain is lengthened, and more cash flow exploration paths are covered, solving the dilemma of being unable to track cash or inter-bank transfers in the current cash flow monitoring process.

[0045] Second, the detection of implicit associations is realized. The present invention provides a solution for detecting implicit capital relationships. In the credit reflux scenario, it can monitor the scenario where the funds eventually flow back to the lender's affiliated enterprises and the loan funds flow back to the lender through the affiliated enterprises, filling the gap in traditional analysis methods.

[0046] Third, the dual-map point and edge attributes can be flexibly defined, and the screening rules can be adjusted through the parameter control module, providing a flexible interface for subsequent adjustment of business objectives and improvement of monitoring accuracy.

[0047] Fourth, through the parameter control module, the exit rules of iterative calculation can be flexibly set according to the system computing resources to achieve a balance between business goals and computing resources.

[0048] Fifth, the proposed idea of ​​multi-graph fusion computing can not only be applied in the field of capital flow monitoring, but also has reference significance for other chain relationship exploration fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0050] Figure 1 It is a flow chart of a method for monitoring capital flow based on dual-graph fusion calculation in an embodiment of the present invention;

[0051] Figure 2 is a flow chart of step 100 in an embodiment of the present invention;

[0052] Figure 3 is a flow chart of step 200 in an embodiment of the present invention;

[0053] Figure 4 is a flow chart of step 300 in an embodiment of the present invention;

[0054] Figure 5 It is a block diagram of a fund flow monitoring system based on dual-graph fusion calculation in a specific embodiment of the present invention;

[0055] Figure 6 A schematic diagram of a process for constructing a customer association relationship map in a specific implementation manner of the present invention;

[0056] Figure 7 A schematic diagram of the process of constructing a capital flow time sequence graph in a specific implementation manner of the present invention;

[0057] Figure 8 It is a schematic diagram of the flow of funds calculation in a specific implementation manner of the present invention;

[0058] Fig. 9 It is a schematic diagram of parameter control flow in a specific implementation manner of the present invention;

[0059] Fig.10 It is a block diagram of a fund flow monitoring device based on dual-graph fusion calculation in an embodiment of the present invention;

[0060] Fig.11 It is a block diagram of the customer relationship map generation module 10 in a specific embodiment of the present invention;

[0061] Fig.12It is a block diagram of the capital flow graph generation module 20 in a specific implementation manner of the present invention;

[0062] Fig.13 It is a block diagram of the capital flow monitoring module 30 in a specific embodiment of the present invention;

[0063] Fig.14 It is a schematic diagram of the structure of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution 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 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 are within the scope of protection of the present invention.

[0065] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0066] It should be noted that the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or apparatuses.

[0067] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0068] The embodiment of the present invention provides a specific implementation method of a fund flow monitoring method based on dual-graph fusion calculation, see Figure 1 , the method specifically includes the following contents:

[0069] Step 100: Generate a customer association relationship map of the target customer based on the target customer ID and the customer association relationship of the target customer.

[0070] It is understandable that the knowledge graph, also known as knowledge domain visualization or knowledge domain mapping map, is a series of various graphics that show the development process and structural relationship of knowledge, using visualization technology to describe knowledge resources and their carriers, and to mine, analyze, construct, draw and display knowledge and their mutual connections. The knowledge graph is a modern theory that combines the theories and methods of applied mathematics, graphics, information visualization technology, information science and other disciplines with metrology citation analysis, co-occurrence analysis and other methods, and uses visualized graphs to vividly display the core structure, development history, frontier fields and overall knowledge architecture of the discipline to achieve the purpose of multidisciplinary integration.

[0071] Step 200: Generate a capital flow time series graph according to the capital flow and capital relationship of the target customer;

[0072] In step 100 and step 200, the target financial customer entity and all entities associated with the target entity are first collected, and then a dual graph is constructed according to the relationship attribute types between the entities, namely, the capital flow time series (dynamic relationship) knowledge graph and the customer association relationship (static relationship) knowledge graph.

[0073] Step 300: Monitor the target customer's fund flow based on the customer relationship graph and the fund flow time series graph.

[0074] Based on the initial directed relationship and exploration target, the relationship connection in the dual graph is calculated according to the preset rules, the qualified capital flow path is obtained, and the monitoring and early warning data is generated;

[0075] From the above description, it can be seen that the embodiment of the present invention provides a method for monitoring capital flow based on dual-graph fusion calculation. First, a customer relationship graph of the target customer is generated based on the target customer ID and the customer relationship of the target customer; then, a capital flow time series graph is generated according to the capital flow direction and capital relationship of the target customer; finally, the capital flow of the target customer is monitored according to the customer relationship graph and the capital flow time series graph. The present invention is based on the method idea of ​​dual-graph fusion calculation, by constructing a customer relationship knowledge graph and a capital flow time series knowledge graph, and applying graph calculation to carry out dynamic and static relationship fusion exploration, so as to restore the hidden capital flow relationship.

[0076] In one embodiment, see Figure 2 , step 100 comprises:

[0077] Step 101: Determine the target customer and the target customer ID based on the business scenario;

[0078] Specifically, the target customer scope and customer relationship types to be established are clarified according to the business scenarios. Taking the analysis of credit fund flow as an example, the target customer scope is the bank's corporate credit customers, and the customer relationship includes seven major relationships: "credit, equity, card holding status, mortgage, guarantee, corporate executives / legal persons, relatives / spouses".

[0079] Step 102: Use a graph database to perform connectivity detection and loop mining on the target customer and its related customers to determine the customer association relationship;

[0080] Step 103: Generate the customer association relationship map according to the target customer ID, the related customer ID and the customer association relationship.

[0081] Based on the customer relationship generated in step 102, a first-hand (direct) relationship is constructed; connectivity and loop mining are performed through the graph database to form penetrating and derived relationships, and second-hand and multi-hand (complex relationships) are completed; each type of relationship is combined with business experience to complete the strong and weak relationship attribute identification.

[0082] In one embodiment, see Figure 3 , step 200 comprises:

[0083] Step 201: Determine the data range of the target customer based on the monitoring scenario;

[0084] The transaction data scope is clarified according to the monitoring scenario. Taking the monitoring of credit fund flow as an example, it is necessary to extract the full transaction flow of loan fund transactions and a certain period of time after the loan is issued. Combined with the clear transaction data scope, the basic data is collected from the business system. In addition to transaction records, the collection scope must also include customer information, account card information, credit ledger, etc., for data integration processing in subsequent point-to-point processing.

[0085] Step 202: Classifying the nodes in the capital flow time series graph according to the counterparty category of the target customer within the data range;

[0086] Step 203: marking the edges in the capital flow time series graph formed by normal capital flow;

[0087] In step 202 and step 203, the same-direction processing of capital flow is first performed: in the transaction record mode of financial institutions, there are at least two modes: unilateral record and bilateral record, that is, the same transaction has two modes of undirected connection (reflecting the lending relationship) or directed connection (inflow and outflow). In this module, it will be uniformly processed into a unidirectional arrow relationship of capital outflow.

[0088] Same entity identification processing for transaction counterparties: In the original transaction records, the transaction counterparty is represented by an account number, card number or customer name. Here, according to the processing method of step 102, the transaction counterparty is processed as a unique entity ID. After processing, the entity of the transaction flow is the same customer ID. Another goal of transaction data processing is to reduce the search space as much as possible, that is, starting from reducing entities and edges, the following processing method is used to identify vertex and edge attributes.

[0089] Processing of public and private identification of counterparties: This identification can be used to determine whether to conduct the next round of relationship exploration and calculation. If the funds flow to an individual, a false connection may be caused due to the same name, and the calculation will be stopped according to this identification. Intra-bank transfers can be defined by existing rules, and for transactions where the counterparty is outside the bank, FastText is used to build a model for classification. FastText is a fast text classifier that provides a simple and efficient text classification and representation method. Based on the algorithm's adaptability to short text classification and fast speed, after using the names of public and private transaction accounts within the bank as samples for model training in this scenario, the model is applied to process counterparties in cross-bank transactions.

[0090] Special point identification processing: When a corporate customer transfers funds to an internal account, a special account or an institutional account, and then transfers funds out of the account, the established capital link is actually a false connection, which will affect the monitoring accuracy. In order to reduce the bridging connection of such special customers, such entities need to be identified. First, through the traditional fuzzy matching method, combined with business experience, the name of the counterparty or the summary note information is associated and matched, and the second is to detect influential nodes in the relationship network through the PageRank algorithm. For a small and medium-sized enterprise and upstream and downstream enterprises, there will not be many counterparties in a certain period of time, and the PageRank value of special customers will be much greater than the target customer due to their fund settlement nature. The PageRank algorithm is used in the transaction flow map to identify customers whose PageRank calculation value exceeds the reasonable value X, as a supplement to the traditional fuzzy matching rules, to increase the coverage of special customers.

[0091] Special edge identification processing: In order to further narrow the data space for searching and reduce the link connections caused by normal capital flows, some transactions such as billing, internal accounts, and deductions are identified according to business rules.

[0092] Step 204: Generate the capital flow time series graph based on the classified nodes and the marked edges.

[0093] Based on the classified nodes and marked edges in step 202 and step 203, a capital flow time series graph is constructed with the target customer's basic information as points and the capital relationship as edges.

[0094] In one embodiment, see Figure 4 , step 300 comprises:

[0095] Step 301: Determine the initial edge of the target customer in the capital flow time series graph, and perform the following iterative operations:

[0096] Step 302: judging whether the transaction counterparty is an entity in the capital flow time series graph according to the first multi-degree capital relationship;

[0097] Step 303: If yes, searching the customer relationship of the counterparty in the customer relationship graph;

[0098] Step 304: Determine whether the counterparty corresponding to the counterparty is an entity in the capital flow time series graph according to the counterparty's customer relationship and the second multi-degree capital relationship, until all customers in the customer relationship graph are calculated.

[0099] In step 301 to step 304, firstly, the initial edge is obtained, such as customer A obtaining a loan from the bank; the first-degree fund relationship that meets the conditions is sequentially searched in the fund flow time series map, that is, customer A transfers funds to B; all the associations of B are searched in the customer association relationship map to form a relationship from B to B' (including the natural self-association relationship from B to B), and the second-degree relationship is established; the third-degree fund relationship that meets the conditions is sequentially searched in the fund flow time series map, that is, the funds are transferred from customer B' to C;

[0100] Next, determine whether the capital flow relationship from B' to C is the target entity or target edge, that is, whether C is the target customer to be found, or whether the transaction from B' to C meets the rule for stopping calculation. If so, stop the calculation, otherwise perform the second round of iteration.

[0101] Search for all the associations of C in the customer association graph, form a relationship from C to C' (including the natural self-association relationship from C to C), and complete the establishment of the fourth-degree relationship; sequentially search for the fifth-degree capital relationship that meets the conditions in the capital flow time series graph, that is, the capital flow relationship from customer C' to D; determine whether the capital flow relationship from C' to D is the target entity or target edge. Iterate according to the above steps.

[0102] In a specific implementation, the present invention takes a financial bank as an example to provide a specific implementation of a method for monitoring capital flow based on dual-graph fusion calculation.

[0103] First, the specific embodiment of the present invention also provides a fund flow monitoring system based on dual-graph fusion calculation, see Figure 5 The system includes: a dual-graph construction device, a capital flow calculation device and a parameter control device, specifically:

[0104] The dual-graph construction device is used to collect the target financial customer entity and all entities associated with the target entity, and construct a dual graph according to the relationship attribute type between the entities, namely, the knowledge graph of capital flow time series (dynamic relationship) and the knowledge graph of customer association relationship (static relationship);

[0105] The capital flow calculation device is used to calculate the relationship connection in the dual graph based on the initial directed relationship and the exploration target according to the preset rules, obtain the capital flow path that meets the conditions, and generate monitoring and early warning data;

[0106] The parameter control device acts on the rule engine of the capital flow calculation device. In order to achieve the business monitoring goals and improve the monitoring effect, the connection conditions are further restricted according to the entity attributes (see Table 1) and relationship attributes, and the connection relationships that meet the conditions are screened according to the monitoring goals.

[0107] Table 1: Entity attributes of fund transaction knowledge graph

[0108]

[0109] See also Figure 6 Based on the above-mentioned fund flow monitoring system based on dual-graph fusion calculation, the customer association relationship graph construction process provided in the specific implementation method of the present invention is as follows:

[0110] Step S101: Data acquisition

[0111] According to the business scenario, the target customer scope and customer relationship types to be established are clarified. Taking the analysis of credit fund flow as an example, the target customer scope is the bank's corporate credit customers, and the customer relationship includes seven major relationships: "credit, equity, card holding status, mortgage, guarantee, corporate executives / legal persons, relatives / spouses".

[0112] Combine available internal and external data sources to complete the collection of basic data from various business systems.

[0113] Step S102: Data processing

[0114] Complete the same entity identification process for the same customer within and outside the bank.

[0115] 1. In-bank customers: The in-bank customer identification number is used uniformly as the customer ID, that is, the same customer identification number is regarded as the same entity;

[0116] 2. Non-bank customers: Use the customer name, business registration number, and organization code to link the bank's customer information table. If the link can be uniquely linked, use the bank's customer identification number as the customer ID. If the link cannot be made, create a virtual 15-digit non-repeating number as the customer ID of the non-bank customer with the same customer name.

[0117] Step S103: Building a customer relationship graph

[0118] Based on the association relationship basic table generated in step S102, construct a first-hand (direct) association relationship;

[0119] Through the graph database, the connectivity and loop mining are carried out to form penetration and derivative association relationships, and to complete second-hand and multi-hand (complex association relationships);

[0120] Each type of relationship is identified with strong and weak relationship attributes based on business experience.

[0121] See also Figure 7 , the process of constructing the capital flow time series graph includes:

[0122] Step S201: data acquisition.

[0123] Clarify the scope of transaction data based on the monitoring scenario. Taking the monitoring of credit fund flows as an example, it is necessary to extract the loan fund transactions and the full transaction flow for a certain period of time after the loan is issued.

[0124] Combined with a clear scope of transaction data, basic data is collected from the business system. In addition to transaction records, the collection scope must also include customer information, account card information, credit ledgers, etc., which will be used for data integration in subsequent point-to-point processing.

[0125] Step S202: data processing.

[0126] Same-direction processing of capital flows: In the transaction record mode of financial institutions, there are at least two modes: unilateral record and bilateral record. That is, for the same transaction, there are two modes: undirected connection (reflecting the lending relationship) or directed connection (inflow and outflow). In this module, they will be uniformly processed into a one-way arrow relationship of capital outflow.

[0127] Processing of the same entity identification of the counterparty: In the original transaction record, the counterparty is represented by an account number, card number or customer name. Here, according to the processing method of step S102, the counterparty is processed as a unique entity ID. After processing, the entity to which the transaction flows is the same customer ID.

[0128] Another goal of transaction data processing is to reduce the search space as much as possible, that is, starting from reducing entities and edges, and using the following processing method to identify vertex and edge attributes (see Table 2).

[0129] Table 2: Edge attributes of fund transaction knowledge graph

[0130]

[0131] Processing of public and private identification of counterparties: This identification can be used to determine whether to conduct the next round of relationship exploration and calculation. If the funds flow to an individual, a false connection may be caused due to the same name, and the calculation will be stopped according to this identification. Intra-bank transfers can be defined by existing rules, and for transactions where the counterparty is outside the bank, FastText is used to build a model for classification. FastText is a fast text classifier that provides a simple and efficient text classification and representation method. Based on the algorithm's adaptability to short text classification and fast speed, after using the names of public and private transaction accounts within the bank as samples for model training in this scenario, the model is applied to process counterparties in cross-bank transactions.

[0132] Special point identification processing: When a corporate customer transfers funds to an internal account, a special account or an institutional account, and then transfers funds out of the account, the established capital link is actually a false connection, which will affect the monitoring accuracy. In order to reduce the bridging connection of such special customers, such entities need to be identified. First, through the traditional fuzzy matching method, combined with business experience, the name of the counterparty or the summary note information is associated and matched, and the second is to detect influential nodes in the relationship network through the PageRank algorithm. For a small and medium-sized enterprise and upstream and downstream enterprises, there will not be many counterparties in a certain period of time, and the PageRank value of special customers will be much greater than the target customer due to their fund settlement nature. The PageRank algorithm is used in the transaction flow map to identify customers whose PageRank calculation value exceeds the reasonable value X, as a supplement to the traditional fuzzy matching rules, to increase the coverage of special customers.

[0133] Special edge identification processing: In order to further narrow the data space for searching and reduce the link connections caused by normal capital flows, some transactions such as billing, internal accounts, and deductions are identified according to business rules.

[0134] Step S203: constructing a time series graph of capital flow.

[0135] Based on the transaction basic table generated in step S202, a capital flow time series graph is constructed with the target customer's basic information as points and the capital relationship as edges.

[0136] See also Figure 8 , the calculation process of funds flow is as follows:

[0137] Specifically, based on the known initial directed edges, the exploration is carried out according to the preset exploration targets (such as exploration depth, target entity, and target edge). Take the credit fund reflux monitoring as an example to illustrate, that is, the initial directed edges are known, and the exploration is carried out in the scenario where the funds flow out and return to the borrower.

[0138] Step S301: Obtaining the initial edge, such as customer A obtaining a loan from a bank;

[0139] Step S302: sequentially search the fund flow time series graph for first-degree fund relationships that meet the conditions, that is, customer A transfers funds to customer B;

[0140] Step S303: Search for all the associations of B in the customer association graph, form a relationship from B to B' (including a natural self-association relationship from B to B), and complete the establishment of a second-degree relationship;

[0141] Step S304: sequentially searching for a three-degree fund relationship that meets the conditions in the fund flow time series graph, that is, transferring funds from customer B' to C;

[0142] Step S305: Determine whether the capital flow relationship from B' to C is the target entity or target edge, that is, whether C is the target customer to be found, or whether the transaction from B' to C meets the rule of stopping calculation. If yes, stop the calculation, otherwise perform the second round of iteration.

[0143] Step S306: Search all the associations of C in the customer association graph to form a relationship from C to C' (including a natural self-association relationship from C to C), thus completing the establishment of a four-degree relationship;

[0144] Step S307: sequentially searching for a five-degree fund relationship that meets the conditions in the fund flow time series graph, that is, the fund flow relationship from customer C' to D;

[0145] Step S308: Determine whether the capital flow relationship from C' to D is a target entity or a target edge.

[0146] Iterate as above.

[0147] See also Fig. 9 , the parameter control process is as follows:

[0148] The parameter control process mainly acts on the capital flow calculation module, that is, by setting point and edge attribute screening conditions and iteration termination conditions, the business monitoring goal is achieved. This module mainly acts on the following data processing process.

[0149] Step S401: In the calculation process of the capital flow time series graph, the relationship set to be used is filtered by the transaction amount attribute, the special point edge identification attribute, and the transaction type attribute;

[0150] Step S402: In the calculation process of the customer relationship graph, the relationship set to be used is selected according to the relationship strength attribute of the customer relationship;

[0151] Step S403: In the capital flow relationship connection calculation process, it is determined whether the association relationship is established through the time attribute and the transaction amount attribute;

[0152] Step S404: In the capital flow relationship connection calculation process, determine whether to terminate the calculation by defining the target entity attributes and the target edge attributes.

[0153] From the above description, it can be seen that the embodiment of the present invention provides a method for monitoring capital flow based on dual-graph fusion calculation. First, a customer relationship graph of the target customer is generated based on the target customer ID and the customer relationship of the target customer; then, a capital flow time series graph is generated according to the capital flow direction and capital relationship of the target customer; finally, the capital flow of the target customer is monitored according to the customer relationship graph and the capital flow time series graph. The present invention is based on the method idea of ​​dual-graph fusion calculation, by constructing a customer relationship knowledge graph and a capital flow time series knowledge graph, and applying graph calculation to carry out dynamic and static relationship fusion exploration, so as to restore the hidden capital flow relationship.

[0154] Based on the same inventive concept, the embodiment of the present application also provides a cash flow monitoring device based on dual-graph fusion calculation, which can be used to implement the method described in the above embodiment, such as the following embodiment. Since the principle of solving the problem by the cash flow monitoring device based on dual-graph fusion calculation is similar to the cash flow monitoring method based on dual-graph fusion calculation, the implementation of the cash flow monitoring device based on dual-graph fusion calculation can refer to the implementation of the cash flow monitoring method based on dual-graph fusion calculation, and the repetitions will not be repeated. As used below, the term "unit" or "module" can be a combination of software and / or hardware that implements a predetermined function. Although the system described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.

[0155] The embodiment of the present invention provides a specific implementation of a fund flow monitoring device based on dual-graph fusion calculation that can implement a fund flow monitoring method based on dual-graph fusion calculation, see Fig.10 The fund flow monitoring device based on dual-graph fusion calculation specifically includes the following contents:

[0156] A customer relationship map generation module 10, configured to generate a customer relationship map of a target customer based on a target customer ID and a customer relationship of the target customer;

[0157] A fund flow graph generation module 20 is used to generate a fund flow time series graph according to the fund flow and fund relationship of the target customer;

[0158] The fund flow monitoring module 30 is used to monitor the fund flow of the target customer according to the customer association relationship map and the fund flow time sequence map.

[0159] In one embodiment, see Fig.11 , the customer relationship map generation module 10 includes:

[0160] A customer determination unit 101 is used to determine the target customer and the target customer ID based on a business scenario;

[0161] The association relationship determination unit 102 is used to perform connectivity detection and loop mining on the target customer and its related customers using a graph database to determine the customer association relationship;

[0162] The customer relationship map generating unit 103 is used to generate the customer relationship map according to the target customer ID, the related customer ID and the customer relationship.

[0163] In one embodiment, see Fig.12 The fund flow graph generation module 20 includes:

[0164] A data range determination unit 201, configured to determine the data range of the target customer based on a monitoring scenario;

[0165] A node classification unit 202 is used to classify the nodes in the capital flow time series graph according to the counterparty category of the target customer within the data range;

[0166] An edge identification unit 203, used for identifying edges in the capital flow time series graph formed by normal capital flow;

[0167] The capital flow graph generation unit 204 is used to generate the capital flow time series graph according to the classified nodes and the marked edges.

[0168] In one embodiment, see Fig.13 , the fund flow monitoring module 30 includes:

[0169] The initial edge determination unit 301 is used to determine the initial edge of the target customer in the capital flow time series graph and perform the following iterative operations:

[0170] A transaction counterparty determination unit 302, configured to determine whether the transaction counterparty is an entity in the capital flow time series graph according to the first multi-degree capital relationship;

[0171] A relationship search unit 303 is used to search for the customer relationship of the counterparty in the customer relationship map.

[0172] The counterparty judgment subunit 304 is used to judge whether the counterparty corresponding to the counterparty in the capital flow time series graph is an entity according to the customer relationship of the counterparty and the second multi-degree capital relationship, until all customers in the customer relationship graph are calculated.

[0173] From the above description, it can be seen that the embodiment of the present invention provides a fund flow monitoring device based on dual-graph fusion calculation, which first generates a customer association relationship map of the target customer based on the target customer ID and the customer association relationship of the target customer; then, generates a fund flow time series map based on the fund flow direction and fund relationship of the target customer; finally, monitors the fund flow of the target customer based on the customer association relationship map and the fund flow time series map. The present invention, through the dual-spectrum fusion calculation idea, combines the dynamic relationship that reflects the transaction timing and flow direction with the static relationship that reflects the customer association relationship to form a calculation mode, breaking through the limitation of traditional fund flow relationship monitoring that only focuses on fund transactions, and realizes the exploration of implicit fund relationships. The beneficial effects of the present invention are as follows:

[0174] First, it breaks through the limitation of relying solely on transaction data to monitor cash flows. After integrating customer relationships, the search space is further expanded, the capital chain is lengthened, and more cash flow exploration paths are covered, solving the dilemma of being unable to track cash or inter-bank transfers in the current cash flow monitoring process.

[0175] Second, the detection of implicit associations is realized. The present invention provides a solution for detecting implicit capital relationships. In the credit reflux scenario, it can monitor the scenario where the funds eventually flow back to the lender's affiliated enterprises and the loan funds flow back to the lender through the affiliated enterprises, filling the gap in traditional analysis methods.

[0176] Third, the dual-map point and edge attributes can be flexibly defined, and the screening rules can be adjusted through the parameter control module, providing a flexible interface for subsequent adjustment of business objectives and improvement of monitoring accuracy.

[0177] Fourth, through the parameter control module, the exit rules of iterative calculation can be flexibly set according to the system computing resources to achieve a balance between business goals and computing resources.

[0178] Fifth, the proposed idea of ​​multi-graph fusion computing can not only be applied in the field of capital flow monitoring, but also has reference significance for other chain relationship exploration fields.

[0179] The embodiments of the present application also provide a specific implementation of an electronic device capable of implementing all steps in the fund flow monitoring method based on dual-graph fusion calculation in the above embodiments, see Fig.14 , electronic equipment specifically includes the following:

[0180] Processor (processor) 1201, memory (memory) 1202, communication interface (CommunicationsInterface) 1203 and bus 1204;

[0181] The processor 1201, the memory 1202, and the communication interface 1203 communicate with each other through the bus 1204; the communication interface 1203 is used to realize information transmission between the server device and the client device and other related devices;

[0182] The processor 1201 is used to call the computer program in the memory 1202. When the processor executes the computer program, all the steps in the fund flow monitoring method based on dual-graph fusion calculation in the above embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0183] Step 100: Generate a customer association relationship map of the target customer based on the target customer ID and the customer association relationship of the target customer;

[0184] Step 200: Generate a capital flow time series graph according to the capital flow and capital relationship of the target customer;

[0185] Step 300: Monitor the target customer's fund flow based on the customer relationship graph and the fund flow time series graph.

[0186] The embodiments of the present application also provide a computer-readable storage medium capable of implementing all the steps in the method for monitoring capital flow based on dual-graph fusion calculation in the above embodiment. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, all the steps in the method for monitoring capital flow based on dual-graph fusion calculation in the above embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0187] Step 100: Generate a customer association relationship map of the target customer based on the target customer ID and the customer association relationship of the target customer;

[0188] Step 200: Generate a capital flow time series graph according to the capital flow and capital relationship of the target customer;

[0189] Step 300: Monitor the target customer's fund flow based on the customer relationship graph and the fund flow time series graph.

[0190] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the hardware + program embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0191] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0192] Although the present application provides method operation steps such as embodiments or flow charts, more or fewer operation steps may be included based on conventional or non-creative labor. The order of steps listed in the embodiments is only one way of executing the order of many steps and does not represent the only execution order. When the actual device or client product is executed, it can be executed in the order of the method shown in the embodiments or the drawings or in parallel (for example, in a parallel processor or multi-threaded processing environment).

[0193] For the convenience of description, the above devices are described in various modules according to their functions. Of course, when implementing the embodiments of this specification, the functions of each module can be implemented in the same or more software and / or hardware, or the module implementing the same function can be implemented by a combination of multiple sub-modules or sub-units. The device embodiments described above are only schematic. For example, the division of the units is only 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 devices or units, which can be electrical, mechanical or other forms.

[0194] Those skilled in the art also know that, in addition to implementing the controller in a purely computer-readable program code, the controller can be made to implement the same function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, such a controller can be considered as a hardware component, and the devices for implementing various functions included therein can also be considered as structures within the hardware component. Or even, the devices for implementing various functions can be considered as both software modules for implementing the method and structures within the hardware component.

[0195] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0196] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0197] The present specification embodiments may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present specification embodiments may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0198] Each embodiment in this specification is described in a progressive manner, and the same and similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. In the description of this specification, the description of the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the embodiment of this specification. In this specification, the schematic representation of the above terms does not necessarily target the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, in the absence of contradiction, a person skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0199] The above is only an example of the embodiment of the present specification and is not intended to limit the embodiment of the present specification. For those skilled in the art, the embodiment of the present specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the embodiment of the present specification shall be included in the scope of the claims of the embodiment of the present specification.

Claims

1. A method for monitoring capital flow based on dual-graph fusion calculation, characterized in that: include: Generate a customer association relationship map of the target customer based on the target customer ID and the customer association relationship of the target customer; Generate a funds flow time series graph based on the funds flow and funds relationship of the target customer; Monitor the target customer's capital flow according to the customer relationship graph and the capital flow time sequence graph; generating a customer association relationship map of the target customer based on the target customer ID and the customer association relationship of the target customer; And generating a capital flow time series graph according to the capital flow and capital relationship of the target customer, including: First, the target financial customer entity and all entities associated with the target entity are collected, and then a dual graph is constructed according to the relationship attribute types between the entities, the dual graph includes: a customer association relationship graph and a fund flow time series graph; wherein the customer association relationship is a static relationship, and the fund flow time series is a dynamic relationship; wherein the types of entities include corporate customers and personal customers, and the attributes of corporate customers include: customer name, internal and external bank identification, borrower identification, public and private customer identification, and special point identification; the attributes of personal customers include: customer name, internal and external bank identification, and special point identification; Monitoring the target customer's fund flow according to the customer association relationship map and the fund flow time sequence map includes: Based on the initial directed relationship and the exploration target, the relationship connection in the dual graph is calculated according to the preset rules, and the qualified capital flow path is obtained to generate monitoring and early warning data.

2. The method for monitoring fund flow based on dual-graph fusion calculation as claimed in claim 1, characterized in that: The step of generating a customer association relationship map of the target customer based on the target customer ID and the customer association relationship of the target customer includes: Determine the target customer and the target customer ID based on the business scenario; Using a graph database to perform connectivity detection and loop mining on the target customer and its related customers to determine the customer association relationship; The customer association relationship map is generated according to the target customer ID, the related customer ID and the customer association relationship.

3. The method for monitoring fund flow based on dual-graph fusion calculation as claimed in claim 1, characterized in that: The generating of a capital flow time series graph according to the capital flow and capital relationship of the target customer comprises: Determine the data scope of the target customer based on the monitoring scenario; Within the data range, classifying the nodes in the capital flow time series graph according to the counterparty category of the target customer; Marking the edges in the capital flow time series graph formed by normal capital flow; The fund flow time series graph is generated according to the classified nodes and the marked edges.

4. The method for monitoring fund flow based on dual-graph fusion calculation as claimed in claim 3, characterized in that: Monitoring the target customer's fund flow according to the customer association relationship map and the fund flow time sequence map includes: Determine the initial edge of the target customer in the capital flow time series graph, and perform the following iterative operations: Determining whether the transaction counterparty is an entity in the capital flow time series graph according to the first multi-degree capital relationship; If yes, searching the customer relationship of the counterparty in the customer relationship graph; According to the customer association relationship of the transaction counterparty and the second multi-degree capital relationship, it is determined in the capital flow time series graph whether the transaction counterparty corresponding to the transaction counterparty is an entity, until all customers in the customer association relationship graph are calculated.

5. A fund flow monitoring device based on dual-graph fusion calculation, characterized in that: include: A customer relationship map generation module, used to generate a customer relationship map of the target customer based on the target customer ID and the customer relationship of the target customer; A fund flow graph generation module is used to generate a fund flow time series graph according to the fund flow and fund relationship of the target customer; A fund flow monitoring module, used to monitor the fund flow of the target customer according to the customer association relationship map and the fund flow time sequence map; The customer relationship map generation module and the capital flow map generation module are specifically used for: First, the target financial customer entity and all entities associated with the target entity are collected, and then a dual graph is constructed according to the relationship attribute types between the entities, the dual graph includes: a customer association relationship graph and a fund flow time series graph; wherein the customer association relationship is a static relationship, and the fund flow time series is a dynamic relationship; wherein the types of entities include corporate customers and personal customers, and the attributes of corporate customers include: customer name, internal and external bank identification, borrower identification, public and private customer identification, and special point identification; the attributes of personal customers include: customer name, internal and external bank identification, and special point identification; The fund flow monitoring module is specifically used for: Based on the initial directed relationship and the exploration target, the relationship connection in the dual graph is calculated according to the preset rules, and the qualified capital flow path is obtained to generate monitoring and early warning data.

6. The fund flow monitoring device based on dual-graph fusion calculation as claimed in claim 5, characterized in that: The customer relationship map generation module includes: A customer determination unit, configured to determine the target customer and the target customer ID based on a business scenario; An association relationship determination unit, configured to perform connectivity detection and loop mining on the target customer and its related customers using a graph database to determine the customer association relationship; A customer relationship map generating unit is used to generate the customer relationship map according to the target customer ID, the related customer ID and the customer relationship.

7. The fund flow monitoring device based on dual-graph fusion calculation as claimed in claim 5, characterized in that: The fund flow graph generation module includes: A data range determination unit, configured to determine the data range of the target customer based on a monitoring scenario; A node classification unit, used for classifying nodes in the capital flow time series graph according to the counterparty category of the target customer within the data range; An edge identification unit, used for identifying edges in the capital flow time series graph formed by normal capital flow; The capital flow graph generation unit is used to generate the capital flow time series graph according to the classified nodes and the marked edges.

8. The fund flow monitoring device based on dual-graph fusion calculation as claimed in claim 7, characterized in that: The fund flow monitoring module includes: The initial edge determination unit is used to determine the initial edge of the target customer in the capital flow time series graph and perform the following iterative operations: A transaction counterparty determination unit, configured to determine whether the transaction counterparty is an entity in the capital flow time series graph according to the first multi-degree capital relationship; an association relationship search unit, configured to search for customer association relationships of the transaction counterparty in the customer association relationship map; The counterparty judgment subunit is used to judge whether the counterparty corresponding to the counterparty in the capital flow time series map is an entity according to the customer relationship of the counterparty and the second multi-degree capital relationship, until all customers in the customer relationship map are calculated.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method for monitoring capital flow based on dual-graph fusion calculation as described in any one of claims 1 to 4 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 for monitoring capital flow based on dual-graph fusion calculation as described in any one of claims 1 to 4 are implemented.

Images