Service recommendation method and device based on bank transaction graph

Abstract

The application discloses a service recommendation method and device based on bank account transaction graph, relates to the technical field of intelligent recommendation, and realizes automatic and accurate recommendation of subsequent services for customers, and improves customer experience. The method comprises the following steps: acquiring bank account transaction historical data, extracting customer nodes and bank account transaction event nodes in the bank account transaction historical data according to a preset node type, connecting edges between the nodes according to a preset connection rule, and obtaining a bank account transaction graph; identifying a bank account transaction event node chain with a time of occurrence within a first preset time length in the bank account transaction graph, determining a business intention corresponding to the bank account transaction event node chain, and determining a correlation level corresponding to the bank account transaction event node chain according to the business intention; and when a new bank account transaction event node of a target customer occurs, recommending a target bank account transaction event node for the target customer according to the correlation level and the new bank account transaction event node.

Description

Technical Field

[0001] This invention relates to the field of intelligent recommendation technology, and in particular to a service recommendation method and apparatus based on bank transaction graphs. Background Technology

[0002] Existing banking systems typically provide transaction confirmation after a customer completes a transaction, but lack in-depth analysis of the underlying needs behind the transaction, thus failing to automatically recommend follow-up services to the customer, resulting in a poor customer experience. Summary of the Invention

[0003] In view of this, the present invention provides a service recommendation method and apparatus based on bank transaction graphs, which realizes automatic and accurate recommendation of follow-up services to customers and improves customer experience.

[0004] According to one aspect of the present invention, a service recommendation method based on bank transaction graphs is provided, the method comprising: Obtain historical bank transaction data, extract customer nodes and bank transaction event nodes from the historical bank transaction data according to preset node types, and connect the edges between each node according to preset connection rules to obtain the bank transaction graph. In the bank transaction graph, identify the chain of bank transaction events that occurred within a first preset time period, determine the business intent corresponding to the chain of bank transaction events, and determine the relevance level of the chain of bank transaction events based on the business intent. When a new bank transaction event occurs for a target customer, a target bank transaction event is recommended for the target customer based on the correlation level and the new bank transaction event.

[0005] Preferably, the preset node types include at least: customer node type and bank transaction event node type; The preset connection rules include at least: preset transaction initiation rules, preset duration connection rules, and preset business logic relevance rules; The process involves extracting customer nodes and bank transaction event nodes from the historical bank transaction data according to preset node types, and connecting the edges between the nodes according to preset connection rules to obtain a bank transaction graph, including: Extract the customer node corresponding to the customer node type and the bank transaction event node corresponding to the bank transaction event node type from the bank transaction history data. Based on the preset transaction initiation rules and the bank's historical transaction data, the edge between the customer node and the bank's transaction event node is obtained; Based on the preset duration connection rule, the preset business logic correlation rule, and the bank transaction history data, the edges between the bank transaction event node and other bank transaction event nodes corresponding to the same customer node are obtained. A bank transaction graph is constructed based on customer nodes, bank transaction event nodes, the edges between customer nodes and bank transaction event nodes, and the edges between bank transaction event nodes and other bank transaction event nodes corresponding to the same customer node.

[0006] Preferably, obtaining the edge between the customer node and the bank transaction event node based on the preset transaction initiation rules and the bank's historical transaction data includes: Based on the preset transaction initiation rules and the bank transaction history data, for any customer node, connect it to the bank transaction event node it initiated to obtain the edge between the customer node and the bank transaction event node. The process of obtaining the edges between a bank transaction event node and other bank transaction event nodes corresponding to the same customer node, based on the preset duration connection rules, the preset business logic relevance rules, and the bank transaction history data, includes: Based on the preset duration connection rules and the historical bank transaction data, for any two bank transaction event nodes connected to the same customer node, if the difference in the occurrence time of the two bank transaction event nodes is within a second preset duration, then the two bank transaction event nodes are connected. Based on the preset business logic relevance rules and the historical bank transaction data, for any two bank transaction event nodes connected to the same customer node, if the two bank transaction event nodes meet the preset business logic relevance, then the two bank transaction event nodes are connected, thus obtaining the edges between the bank transaction event node and other bank transaction event nodes corresponding to the same customer node.

[0007] Preferably, identifying the chain of bank transaction event nodes that occurred within a first preset time period in the bank transaction graph includes: For any of the aforementioned customer nodes, sort the multiple bank transaction event nodes connected to it in chronological order, and connect the bank transaction event nodes that occurred within a first preset time period into a bank transaction event node chain.

[0008] Preferably, determining the business intent corresponding to the bank transaction event node chain includes: Based on the edges of the bank transaction graph and the preset business intent type, the business intent corresponding to the bank transaction event node chain is identified.

[0009] Preferably, determining the relevance level of the bank transaction event node chain based on the business intent includes: The relevance level of the bank transaction event node chain is determined based on the number of bank transaction event node chains with the same business intent.

[0010] Preferably, recommending target bank transaction event nodes to the target customer based on the relevance level and the new bank transaction event node includes: Based on all bank transaction event node chains corresponding to all customer nodes, match at least one candidate bank transaction event node chain corresponding to the target customer. In each candidate bank transaction event node chain, locate the matching bank transaction event node that is the same as the new bank transaction event node, and determine the adjacent subsequent bank transaction event nodes of each matching bank transaction event node; Based on the relevance level of all candidate bank transaction event node chains and their corresponding adjacent subsequent bank transaction event nodes, the target bank transaction event node is determined and recommended to the target customer.

[0011] According to another aspect of the present invention, a service recommendation device based on a bank transaction graph is provided, the device comprising: The acquisition module is used to acquire historical data of bank transactions, extract customer nodes and bank transaction event nodes from the historical data of bank transactions according to preset node types, and connect the edges between each node according to preset connection rules to obtain the bank transaction graph. The determination module is used to identify a chain of bank transaction event nodes that occurred within a first preset time period in the bank transaction graph, determine the business intent corresponding to the chain of bank transaction event nodes, and determine the relevance level corresponding to the chain of bank transaction event nodes based on the business intent. The recommendation module is used to recommend target bank transaction events to the target customer when a new bank transaction event occurs, based on the relevance level and the new bank transaction event.

[0012] According to another aspect of the present invention, a storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the above-described service recommendation method based on bank transaction graphs.

[0013] According to another aspect of the present invention, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor, when executing the program, implements the above-described service recommendation method based on bank transaction graphs.

[0014] By employing the above technical solutions, this invention provides a service recommendation method and apparatus based on a bank transaction graph. Through this invention, a bank transaction graph is constructed, and the relationship between customer nodes and bank transaction event nodes is structurally represented. This allows for further mining of at least one bank transaction event node chain corresponding to each customer node, and determination of relevance levels based on business intent. This improves the accuracy of transaction association analysis. When a new bank transaction event node occurs for a target customer, recommendations are made based on the bank transaction event node chain and relevance level. This enables rapid and accurate output of subsequent transactions (i.e., target bank transaction event nodes) that conform to the target customer's habits, achieving personalized and highly reliable intelligent recommendations for bank transactions, thereby improving recommendation efficiency and customer experience.

[0015] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of this application. In the drawings: Figure 1 The illustration shows a flowchart of a service recommendation method based on bank transaction graphs provided by an embodiment of the present invention; Figure 2 This illustration shows a flowchart of another service recommendation method based on bank transaction graphs provided by an embodiment of the present invention; Figure 3 This diagram illustrates the structure of a service recommendation device based on a bank transaction graph according to an embodiment of the present invention. Figure 4 This diagram illustrates the structure of another service recommendation device based on bank transaction graphs provided in an embodiment of the present invention. Detailed Implementation

[0017] The present invention will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0018] This embodiment provides a service recommendation method based on bank transaction graphs, such as... Figure 1 As shown, the method includes: 101. Obtain historical bank transaction data, extract customer nodes and bank transaction event nodes from the historical bank transaction data according to the preset node type, and connect the edges between the nodes according to the preset connection rules to obtain the bank transaction graph.

[0019] It should be noted that bank transaction activity refers to transactions in which funds change in a customer's bank account.

[0020] In order to construct a bank transaction graph, a structured representation of the relationship between customer nodes and bank transaction event nodes is presented.

[0021] 102. Identify the chain of bank transaction events that occurred within a first preset time period in the bank transaction graph, determine the business intent corresponding to the chain of bank transaction events, and determine the relevance level of the chain of bank transaction events based on the business intent.

[0022] In this embodiment, since it is necessary to recommend target bank transaction event nodes to target customers, at least one bank transaction event node chain corresponding to each customer node is constructed based on each customer node. Specifically, a series of bank transaction event nodes within this short time period within the first preset time period are extracted, sorted according to time order, and strung together to form a bank transaction event node chain, so that at least one bank transaction event node chain can be identified.

[0023] 103. When a new bank transaction event occurs for a target customer, recommend a target bank transaction event for the target customer based on the correlation level and the new bank transaction event.

[0024] In this embodiment, the customer node that is the same as the target customer is queried, and the bank transaction event node chain that does not have the new bank transaction event node is removed from all bank transaction event node chains of the same customer node to obtain all candidate bank transaction event node chains of the target customer.

[0025] This invention provides a service recommendation method and apparatus based on a bank transaction graph. The technical solution of this invention constructs a bank transaction graph, structurally representing the relationship between customer nodes and bank transaction event nodes. This allows for further mining of at least one bank transaction event node chain corresponding to each customer node, and determination of relevance levels based on business intent. This improves the accuracy of transaction association analysis. When a new bank transaction event node occurs for a target customer, recommendations are made based on the bank transaction event node chain and relevance level. This enables rapid and accurate output of subsequent transactions (i.e., target bank transaction event nodes) that conform to the target customer's habits, achieving personalized and highly reliable intelligent bank transaction recommendations, thereby improving recommendation efficiency and customer experience.

[0026] Furthermore, as a refinement and extension of the specific implementation methods of the above embodiments, and to fully illustrate the specific implementation process in this embodiment, another service recommendation method based on bank transaction graphs is provided, such as... Figure 2 As shown, the method includes: 201. Obtain historical bank transaction data, extract customer nodes and bank transaction event nodes from the historical bank transaction data according to the preset node type, and connect the edges between each node according to the preset connection rules to obtain the bank transaction graph.

[0027] In this embodiment, the preset node types include at least: customer node type and bank transaction event node type; the preset connection rules include at least: preset transaction initiation rules, preset duration connection rules and preset business logic relevance rules; The step of extracting customer nodes and bank transaction event nodes from the historical bank transaction data according to preset node types, and connecting the edges between the nodes according to preset connection rules to obtain a bank transaction graph, includes: extracting customer nodes corresponding to the customer node type and bank transaction event nodes corresponding to the bank transaction event node type from the historical bank transaction data; obtaining the edges between the customer nodes and bank transaction event nodes based on the preset transaction initiation rules and the historical bank transaction data; obtaining the edges between the bank transaction event nodes and other bank transaction event nodes corresponding to the same customer node based on the preset duration connection rules, the preset business logic relevance rules, and the historical bank transaction data; and constructing a bank transaction graph based on the customer nodes, bank transaction event nodes, the edges between the customer nodes and bank transaction event nodes, and the edges between the bank transaction event nodes and other bank transaction event nodes corresponding to the same customer node.

[0028] The step of obtaining the edge between the customer node and the bank transaction event node based on the preset transaction initiation rule and the bank transaction history data includes: for any customer node, connecting it to the bank transaction event node it initiated, based on the preset transaction initiation rule and the bank transaction history data, to obtain the edge between the customer node and the bank transaction event node; the step of obtaining the edge between the bank transaction event node and other bank transaction event nodes corresponding to the same customer node based on the preset duration connection rule, the preset business logic relevance rule, and the bank transaction history data includes: based on the preset duration connection... Based on the rules and the historical bank transaction data, for any two bank transaction event nodes connected to the same customer node, if the difference in the occurrence time of the two bank transaction event nodes is within a second preset time period, then the two bank transaction event nodes are connected. Based on the preset business logic relevance rules and the historical bank transaction data, for any two bank transaction event nodes connected to the same customer node, if the two bank transaction event nodes meet the preset business logic relevance, then the two bank transaction event nodes are connected, thus obtaining the edges between the bank transaction event node and other bank transaction event nodes corresponding to the same customer node.

[0029] This involves obtaining a unique identifier set for each customer, with each unique identifier corresponding to one customer node.

[0030] Bank transaction event nodes refer to bank transaction-related businesses, including retail customer transaction types and corporate customer transaction types. For retail customer transaction types, for example: (1) deposits and withdrawals: including counter cash deposits and withdrawals, ATM deposits and withdrawals, and smart counter deposits and withdrawals; (2) transfers and remittances: intra-bank transfers (can be to one's own or another person's account), inter-bank transfers (pay attention to the channel), and cross-border remittances; (3) consumer payments: card payments, POS payments, online payments, and direct debit; (4) investment and wealth management: purchasing funds, wealth management products, government bonds, redeeming wealth management products upon maturity, and receiving fund dividends; (5) credit-related: loan disbursements, loan repayments, credit card overdrafts, and credit card repayments. For corporate clients, the types of transactions include: (1) Settlement: transfers, remittances, and bill settlements between corporate accounts; (2) Payment and receipt: receiving payments from clients, paying payments to suppliers, distributing employee wages, and paying taxes; (3) Financing: bank loans disbursed, repayment of loan principal and interest, and payment of acceptance bill margin; (4) Other business: account management fees, handling fee deductions, and changes in funds related to investment business.

[0031] The preset transaction initiation rule means that if a client node initiates a bank transaction event node, a connection relationship is established between the client node and the bank transaction event node, thus forming an edge between the client node and the bank transaction event node.

[0032] The preset duration connection rule means that if a client node initiates two different bank transaction event nodes at two different times, that is, if the occurrence time of these two different bank transaction event nodes is different, and the difference in occurrence time is within the second preset duration, then a connection relationship is established between these two different bank transaction event nodes, thus forming an edge between the two different bank transaction event nodes.

[0033] It should be noted that step 202 of the embodiment uses a first preset duration. The difference between the first preset duration and the second preset duration is that the first preset duration is used to divide all bank transaction event nodes connected to (that is, initiated) by each customer node, and the second preset duration is used to determine whether a connection relationship is established between two different bank transaction event nodes connected to (that is, initiated) by each customer node. The first preset duration is greater than or equal to the second preset duration.

[0034] The pre-defined business logic relevance rule means that if two different bank transaction events initiated by a certain customer node are related by one of the pre-defined business logic relevance rules, then a connection is established between these two different bank transaction events, forming an edge between them. For example, if the two different bank transaction events are a wealth management product redemption upon maturity and a credit card repayment, since wealth management product redemption upon maturity and credit card repayment are among the pre-defined business logic relevance rules, a connection is established between these two different bank transaction events.

[0035] Among them, all preset business logic relevance is set in advance according to the following three business logics: Business Logic (1) Fund allocation mode with debt repayment as the core. For example, redeeming low-liquidity assets to repay short-term liabilities, specifically, redeeming fixed-term wealth management products, funds, etc. to repay credit card debts. For example, splitting multiple assets for redemption to cover large repayments. This reflects the customer's habit of using funds. Business Logic (2) Fund allocation mode with asset optimization as the core. For example, redeeming low-yield, high-risk products and investing in higher-yield products. For example, redeeming closed-end products and transferring them to flexible deposit and withdrawal products. For example, migrating assets across platforms and categories, specifically, redeeming wealth management products from Bank A and transferring them to structured deposits from Bank B. Generally affected by promotional activities, platform preferences, etc., reflecting the customer's sensitivity to cost-effectiveness. Business Logic (3) Emergency mode with temporary fund turnover as the core. For example, redeeming assets that can be redeemed at any time. For example, temporary fund turnover, real-time and without regard to loss.

[0036] For example, if 10 preset business logic related items are set according to business logic (1), 20 preset business logic related items are set according to business logic (2), and 20 preset business logic related items are set according to business logic (3), then there are a total of 50 preset business logic related items.

[0037] Preferably, the preset node type further includes: transaction object node type, and the preset connection rule further includes: preset transaction-related rules. Transaction object nodes are extracted from the bank's historical transaction data according to the preset node type. Specifically, transaction object nodes corresponding to the transaction object node type are extracted from the bank's historical transaction data. The transaction object node type includes individuals, merchants, institutions, specific accounts, etc. The step of connecting the edges between nodes according to the preset connection rules to obtain the bank's transaction graph further includes: based on the preset transaction-related rules and the bank's historical transaction data, for any bank transaction event node, connecting it with its related transaction object nodes to obtain the edges between the bank transaction event node and the transaction object nodes.

[0038] 202. Identify the chain of bank transaction event nodes that occurred within a first preset time period in the bank transaction graph.

[0039] In this embodiment, identifying the chain of bank transaction event nodes that occurred within a first preset duration in the bank transaction graph includes: for any customer node, sorting its connected multiple bank transaction event nodes in chronological order, and connecting each bank transaction event node that occurred within the first preset duration into a chain of bank transaction event nodes.

[0040] This involves extracting all bank transaction event nodes connected to (i.e. initiated) by each customer node, sorting them according to the time sequence from the earliest occurrence time to the latest occurrence time, and truncating them from the earliest occurrence time to the latest occurrence time using a first preset duration as the standard. For example, if the difference between the earliest occurrence time and the occurrence time of a certain bank transaction event node (named the first occurrence time) is less than or equal to a first preset duration, then bank transaction event nodes that are greater than or equal to the earliest occurrence time to less than or equal to the first occurrence time are connected in chronological order to obtain the first bank transaction event node chain. Next, the occurrence times of the adjacent subsequent bank transaction event nodes of the first occurrence time are obtained (named the second occurrence time). If the difference between the second occurrence time and the occurrence time of a certain bank transaction event node (named the third occurrence time) is less than or equal to the first preset duration, then bank transaction event nodes that are greater than or equal to the second occurrence time to less than or equal to the third occurrence time are connected in chronological order to obtain the second bank transaction event node chain. This process continues until the latest occurrence time, resulting in the last bank transaction event node chain.

[0041] For example, the 10 bank transaction event nodes connected to customer node A are ordered chronologically as follows: Bank Transaction Event Node 1 (corresponding to the earliest occurrence time), Bank Transaction Event Node 2, Bank Transaction Event Node 3, Bank Transaction Event Node 4, Bank Transaction Event Node 5, Bank Transaction Event Node 6, Bank Transaction Event Node 7, Bank Transaction Event Node 8, Bank Transaction Event Node 9, and Bank Transaction Event Node 10 (corresponding to the latest occurrence time). Where the difference between the earliest occurrence time and the occurrence time of Bank Transaction Event Node 3 is less than or equal to a first preset duration, then the occurrence time of Bank Transaction Event Node 3 is designated as the first occurrence time. The first bank transaction event node chain, namely bank transaction event nodes 1, 2, and 3, is obtained. The occurrence time of bank transaction event node 4 is named the second occurrence time. The difference between the second occurrence time and the occurrence time of bank transaction event node 8 is less than or equal to the first preset duration. Then, the occurrence time of bank transaction event node 8 is named the third occurrence time, thus obtaining the second bank transaction event node chain, namely bank transaction event nodes 4, 5, 6, 7, and 8. The occurrence time of bank transaction event node 9 is named the fourth occurrence time. The difference between the fourth occurrence time and the latest occurrence time is less than or equal to the first preset duration, thus obtaining the third bank transaction event node chain, namely bank transaction event nodes 9 and 10.

[0042] 203. Identify the business intent corresponding to the bank transaction event node chain based on the edges of the bank transaction graph and the preset business intent type.

[0043] In this embodiment, for any customer node, the edges between it and bank transaction event nodes, the edges between two different bank transaction event nodes connected to it, and the edges between a bank transaction event node connected to it and the transaction object node are extracted through the bank transaction graph to obtain the sub-bank transaction graph corresponding to the customer node. Based on the sub-bank transaction graph corresponding to the customer node, the business intent corresponding to the bank transaction event node chain is identified according to the following three preset business intent types.

[0044] (1) Preset business intent type: repetitive pattern, identify whether it is a repetitive pattern business intent, such as periodic repayment or fixed investment; (2) Preset business intent type: fixed sequence pattern, identify whether it is a fixed sequence pattern business intent, such as each sequence (occurring at a non-fixed time) is a fixed number of bank transaction event nodes, occurring in a fixed order, such as transferring a small amount from bank A to bank B at a non-fixed time, accumulating a small amount in bank B to a certain amount of assets, and buying a certain product of bank B. Its business intent is that the product of bank B has a fixed capital investment threshold, and accumulating a small amount to reach the fixed capital investment threshold; (3) Preset business intent type: fixed transaction object node, identify whether it is a fixed transaction object node business intent.

[0045] It should be noted that when identifying the business intent corresponding to each bank transaction event node chain, the specific business intent in the preset business intent type is output. For example, for a certain customer node: the business intent output by bank transaction event node chain 1 is periodic repayment, the business intent output by bank transaction event node chain 2 is fixed investment, the business intent output by bank transaction event node chain 3 is fixed sequence pattern 1, the business intent output by bank transaction event node chains 4 and 5 is fixed sequence pattern 2, the business intent output by bank transaction event node chain 6 is fixed transaction object node 1, and the business intent output by bank transaction event node chains 7, 8, 9, and 10 is fixed transaction object node 2.

[0046] 204. Determine the relevance level of the bank transaction event node chain based on the number of bank transaction event node chains with the same business intent.

[0047] In this embodiment, for any customer node, there is at least one bank transaction event node chain. In step 203 of the above embodiment, the corresponding business intent is identified for all the bank transaction event node chains (a total of 10) included for the customer node. The number of bank transaction event node chains with the same business intent for the customer node is calculated and sorted in descending order of the number to determine the relevance level of its bank transaction event node chain.

[0048] As one implementation method, the relevance level is set to three levels: absolutely trustworthy, trustworthy, and generally trustworthy. The absolutely trustworthy level is the most numerous, which is the bank transaction event node chain 7, 8, 9, and 10 corresponding to the above-mentioned fixed transaction object node 2. The trustworthy level is the next most numerous, which is the bank transaction event node chain 4 and 5 corresponding to the above-mentioned fixed sequence pattern 2. The generally trustworthy level is the least numerous, which is the above-mentioned bank transaction event node chain 1, 2, 3, and 6.

[0049] 205. When a new bank transaction event occurs for a target customer, recommend a target bank transaction event for the target customer based on the correlation level and the new bank transaction event.

[0050] In this embodiment, recommending a target bank transaction event node to the target customer based on the relevance level and the new bank transaction event node includes: matching at least one candidate bank transaction event node chain corresponding to the target customer based on all bank transaction event node chains corresponding to all customer nodes; locating a matching bank transaction event node that is the same as the new bank transaction event node in each candidate bank transaction event node chain, determining the adjacent subsequent bank transaction event nodes of each matching bank transaction event node; and determining the target bank transaction event node based on the relevance level corresponding to all candidate bank transaction event node chains and the corresponding adjacent subsequent bank transaction event nodes, and recommending it to the target customer.

[0051] Specifically, the process involves querying customer nodes that are identical to the target customer, removing bank transaction event node chains that do not contain the new bank transaction event node from all bank transaction event node chains of the same customer node, obtaining all candidate bank transaction event node chains for the target customer, locating matching bank transaction event nodes that are identical to the new bank transaction event node in each candidate bank transaction event node chain, determining the adjacent subsequent bank transaction event nodes for each matching bank transaction event node, and assigning the relevance level corresponding to the candidate bank transaction event node chain to the corresponding adjacent subsequent bank transaction event node.

[0052] As one implementation method, the target bank transaction event node is identified as the one with the highest relevance among all adjacent subsequent bank transaction event nodes and recommended to the target customer. It should be noted that there must be at least one adjacent subsequent bank transaction event node with the highest relevance; that is, there must be at least one target bank transaction event node.

[0053] This invention provides a service recommendation method and apparatus based on a bank transaction graph. The technical solution of this invention constructs a bank transaction graph, structurally representing the relationship between customer nodes and bank transaction event nodes. This allows for further mining of at least one bank transaction event node chain corresponding to each customer node, and determination of relevance levels based on business intent. This improves the accuracy of transaction association analysis. When a new bank transaction event node occurs for a target customer, recommendations are made based on the bank transaction event node chain and relevance level. This enables rapid and accurate output of subsequent transactions (i.e., target bank transaction event nodes) that conform to the target customer's habits, achieving personalized and highly reliable intelligent bank transaction recommendations, thereby improving recommendation efficiency and customer experience.

[0054] Furthermore, as Figure 1 and Figure 2 The specific implementation of the method shown in this invention provides a service recommendation device based on a bank transaction graph, such as... Figure 3 As shown, the device includes: an acquisition module 31, a determination module 32, and a recommendation module 33; The acquisition module 31 is used to acquire historical data of bank transactions, extract customer nodes and bank transaction event nodes from the historical data of bank transactions according to preset node types, and connect the edges between each node according to preset connection rules to obtain a bank transaction graph. The determination module 32 is used to identify a chain of bank transaction event nodes that occurred within a first preset time period in the bank transaction graph, determine the business intent corresponding to the chain of bank transaction event nodes, and determine the relevance level corresponding to the chain of bank transaction event nodes based on the business intent. Recommendation module 33 is used to recommend target bank transaction events to the target customer when a new bank transaction event occurs, based on the relevance level and the new bank transaction event.

[0055] Accordingly, the preset node types include at least: customer node types and bank transaction event node types; the preset connection rules include at least: preset transaction initiation rules, preset duration connection rules, and preset business logic relevance rules; in order to extract customer nodes and bank transaction event nodes from the historical bank transaction data according to the preset node types, and to connect the edges between each node according to the preset connection rules to obtain a bank transaction graph, the acquisition module 31 is used to extract the customer nodes corresponding to the customer node types and the bank transaction event nodes corresponding to the bank transaction event node types from the historical bank transaction data; based on The preset transaction initiation rules and the historical bank transaction data are used to obtain the edges between the customer node and the bank transaction event node; based on the preset duration connection rules, the preset business logic relevance rules, and the historical bank transaction data, the edges between the bank transaction event node and other bank transaction event nodes corresponding to the same customer node are obtained; and a bank transaction graph is constructed based on the customer node, the bank transaction event node, the edges between the customer node and the bank transaction event node, and the edges between the bank transaction event node and other bank transaction event nodes corresponding to the same customer node.

[0056] Accordingly, in order to obtain the edge between the customer node and the bank transaction event node based on the preset transaction initiation rules and the bank transaction history data, the acquisition module 31 is used to connect any customer node to the bank transaction event node it initiated, based on the preset transaction initiation rules and the bank transaction history data, to obtain the edge between the customer node and the bank transaction event node; in order to obtain the edge between the bank transaction event node and other bank transaction event nodes corresponding to the same customer node based on the preset duration connection rules, the preset business logic relevance rules, and the bank transaction history data, the acquisition module 31 is used to Based on the preset duration connection rules and the historical bank transaction data, for any two bank transaction event nodes connected to the same customer node, if the difference in the occurrence time of the two bank transaction event nodes is within a second preset duration, then the two bank transaction event nodes are connected. Based on the preset business logic relevance rules and the historical bank transaction data, for any two bank transaction event nodes connected to the same customer node, if the two bank transaction event nodes meet the preset business logic relevance, then the two bank transaction event nodes are connected, thus obtaining the edges between the bank transaction event node and other bank transaction event nodes corresponding to the same customer node.

[0057] Accordingly, in order to identify the chain of bank transaction event nodes that occurred within a first preset duration in the bank transaction graph, the determining module 32 is used to sort the multiple bank transaction event nodes connected to any customer node in chronological order, and connect each bank transaction event node that occurred within the first preset duration into a chain of bank transaction event nodes.

[0058] Accordingly, in order to determine the business intent corresponding to the bank transaction event node chain, the determining module 32 is used to identify the business intent corresponding to the bank transaction event node chain based on the edges of the bank transaction graph and the preset business intent type.

[0059] Accordingly, in order to determine the relevance level of the bank transaction event node chain based on the business intent, the determining module 32 is used to determine the relevance level of the bank transaction event node chain based on the business intent and the number of bank transaction event node chains with the same business intent.

[0060] Accordingly, in order to recommend target bank transaction event nodes to the target customer based on the relevance level and the new bank transaction event node, the recommendation module 33 includes: a matching unit 331 and a determining unit 332; The matching unit 331 is used to match at least one candidate bank transaction event node chain corresponding to the target customer based on all bank transaction event node chains corresponding to all customer nodes. The determining unit 332 is used to locate the matching bank transaction event node that is the same as the new bank transaction event node in each candidate bank transaction event node chain, determine the adjacent subsequent bank transaction event nodes of each matching bank transaction event node, and determine the target bank transaction event node based on the relevance level of all candidate bank transaction event node chains and the corresponding adjacent subsequent bank transaction event nodes, and recommend it to the target customer.

[0061] It should be noted that other corresponding descriptions of the functional units involved in the service recommendation device based on bank transaction graphs provided in this embodiment can be found in [reference needed]. Figures 1 to 2 The corresponding description will not be repeated here.

[0062] Based on the above, Figures 1 to 2 Accordingly, this embodiment also provides a storage medium, which may be volatile or non-volatile, storing a computer program that, when executed by a processor, implements the above-described method. Figures 1 to 2 The service recommendation method shown is based on the bank's transaction graph.

[0063] Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, portable hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of the present invention.

[0064] Based on the above, Figures 1 to 2 The method shown and Figure 3 , Figure 4 To achieve the above objectives, the present application also provides a computer device, specifically a personal computer, server, network device, etc., as shown in the illustrated embodiment. This computer device includes a storage medium and a processor; the storage medium stores a computer program; the processor executes the computer program to achieve the above-described objectives. Figure 1 and Figure 2 The service recommendation method shown is based on the bank's transaction graph.

[0065] Optionally, the computer device may also include a client interface, a network interface, a camera, radio frequency (RF) circuitry, sensors, audio circuitry, a Wi-Fi module, etc. The client interface may include a display screen, input units such as a keyboard, etc., and optional client interfaces may also include USB interfaces, card reader interfaces, etc. The network interface may optionally include standard wired interfaces, wireless interfaces (such as Wi-Fi interfaces), etc.

[0066] Those skilled in the art will understand that the computer device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0067] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the aforementioned computer device, supporting the operation of information processing programs and other software and / or programs. The network communication module is used to enable communication between the various components within the non-volatile storage medium, as well as communication with other hardware and software in the information processing entity device.

[0068] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general-purpose hardware platform, or it can be implemented by hardware.

[0069] This invention provides a service recommendation method and apparatus based on a bank transaction graph. The technical solution of this invention constructs a bank transaction graph, structurally representing the relationship between customer nodes and bank transaction event nodes. This allows for further mining of at least one bank transaction event node chain corresponding to each customer node, and determination of relevance levels based on business intent. This improves the accuracy of transaction association analysis. When a new bank transaction event node occurs for a target customer, recommendations are made based on the bank transaction event node chain and relevance level. This enables rapid and accurate output of subsequent transactions (i.e., target bank transaction event nodes) that conform to the target customer's habits, achieving personalized and highly reliable intelligent bank transaction recommendations, thereby improving recommendation efficiency and customer experience.

[0070] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing the present invention. Those skilled in the art will understand that the modules in the apparatus of the embodiment can be distributed within the apparatus of the embodiment as described, or they can be located in one or more apparatuses different from this embodiment, with corresponding changes. The modules of the above-described embodiment can be combined into one module, or further divided into multiple sub-modules.

[0071] The serial numbers used above are for descriptive purposes only and do not represent the superiority or inferiority of any particular implementation scenario. The above disclosures are merely a few specific implementation scenarios of the present invention; however, the present invention is not limited thereto, and any variations conceived by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. A service recommendation method based on bank transaction graphs, characterized in that, The method includes: Obtain historical bank transaction data, extract customer nodes and bank transaction event nodes from the historical bank transaction data according to preset node types, and connect the edges between each node according to preset connection rules to obtain the bank transaction graph. In the bank transaction graph, identify the chain of bank transaction events that occurred within a first preset time period, determine the business intent corresponding to the chain of bank transaction events, and determine the relevance level of the chain of bank transaction events based on the business intent. When a new bank transaction event occurs for a target customer, a target bank transaction event is recommended for the target customer based on the correlation level and the new bank transaction event.

2. The method according to claim 1, characterized in that, The preset node types include at least: customer node type and bank transaction event node type; The preset connection rules include at least: preset transaction initiation rules, preset duration connection rules, and preset business logic relevance rules; The process involves extracting customer nodes and bank transaction event nodes from the historical bank transaction data according to preset node types, and connecting the edges between the nodes according to preset connection rules to obtain a bank transaction graph, including: Extract the customer node corresponding to the customer node type and the bank transaction event node corresponding to the bank transaction event node type from the bank transaction history data. Based on the preset transaction initiation rules and the bank's historical transaction data, the edge between the customer node and the bank's transaction event node is obtained; Based on the preset duration connection rule, the preset business logic correlation rule, and the bank transaction history data, the edges between the bank transaction event node and other bank transaction event nodes corresponding to the same customer node are obtained. A bank transaction graph is constructed based on customer nodes, bank transaction event nodes, the edges between customer nodes and bank transaction event nodes, and the edges between bank transaction event nodes and other bank transaction event nodes corresponding to the same customer node.

3. The method according to claim 2, characterized in that, The step of obtaining the edge between the customer node and the bank transaction event node based on the preset transaction initiation rules and the bank's historical transaction data includes: Based on the preset transaction initiation rules and the bank transaction history data, for any customer node, connect it to the bank transaction event node it initiated to obtain the edge between the customer node and the bank transaction event node. The process of obtaining the edges between a bank transaction event node and other bank transaction event nodes corresponding to the same customer node, based on the preset duration connection rules, the preset business logic relevance rules, and the bank transaction history data, includes: Based on the preset duration connection rules and the historical bank transaction data, for any two bank transaction event nodes connected to the same customer node, if the difference in the occurrence time of the two bank transaction event nodes is within a second preset duration, then the two bank transaction event nodes are connected. Based on the preset business logic relevance rules and the historical bank transaction data, for any two bank transaction event nodes connected to the same customer node, if the two bank transaction event nodes meet the preset business logic relevance, then the two bank transaction event nodes are connected, thus obtaining the edges between the bank transaction event node and other bank transaction event nodes corresponding to the same customer node.

4. The method according to claim 2 or 3, characterized in that, The step of identifying the chain of bank transaction event nodes that occurred within a first preset time period in the bank transaction graph includes: For any of the aforementioned customer nodes, sort the multiple bank transaction event nodes connected to it in chronological order, and connect the bank transaction event nodes that occurred within a first preset time period into a bank transaction event node chain.

5. The method according to claim 1, characterized in that, The process of determining the business intent corresponding to the bank's transaction event node chain includes: Based on the edges of the bank transaction graph and the preset business intent type, the business intent corresponding to the bank transaction event node chain is identified.

6. The method according to claim 1, characterized in that, The step of determining the relevance level of the bank transaction event node chain according to the business intent includes: The relevance level of the bank transaction event node chain is determined based on the number of bank transaction event node chains with the same business intent.

7. The method according to claim 1, characterized in that, The step of recommending target bank transaction event nodes to the target customer based on the correlation level and the new bank transaction event node includes: Based on all bank transaction event node chains corresponding to all customer nodes, match at least one candidate bank transaction event node chain corresponding to the target customer. In each candidate bank transaction event node chain, locate the matching bank transaction event node that is the same as the new bank transaction event node, and determine the adjacent subsequent bank transaction event nodes of each matching bank transaction event node; Based on the relevance level of all candidate bank transaction event node chains and their corresponding adjacent subsequent bank transaction event nodes, the target bank transaction event node is determined and recommended to the target customer.

8. A service recommendation device based on bank transaction activity graphs, characterized in that, The device includes: The acquisition module is used to acquire historical data of bank transactions, extract customer nodes and bank transaction event nodes from the historical data of bank transactions according to preset node types, and connect the edges between each node according to preset connection rules to obtain the bank transaction graph. The determination module is used to identify a chain of bank transaction event nodes that occurred within a first preset time period in the bank transaction graph, determine the business intent corresponding to the chain of bank transaction event nodes, and determine the relevance level corresponding to the chain of bank transaction event nodes based on the business intent. The recommendation module is used to recommend target bank transaction events to the target customer when a new bank transaction event occurs, based on the relevance level and the new bank transaction event.

9. A storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the service recommendation method based on the bank transaction graph as described in any one of claims 1 to 7.

10. A computer device comprising a memory, a processor, and a computer program stored on a storage medium and executable on the processor, characterized in that, When the processor executes the program, it implements the service recommendation method based on bank transaction graphs as described in any one of claims 1 to 7.

Images