Cross-border clearing path recommendation method and system, electronic equipment and storage medium

Abstract

The invention provides a cross-border liquidation path recommendation method and system, electronic equipment and a storage medium, and relates to the technical field of financial science and technology, and the method comprises the steps: constructing a cross-border liquidation path knowledge graph according to the data of a plurality of data sources, the cross-border liquidation path knowledge graph comprises a liquidation path relation, a liquidation cost relation and a liquidation timeliness relation among bank nodes; a target cross-border clearing path is determined from the cross-border clearing path knowledge graph according to the cross-border remittance information, the target cross-border clearing path is visually displayed, a head node of the target cross-border clearing path is a remittance bank, and a tail node of the target cross-border clearing path is a collection bank in the cross-border remittance information. The clearing currencies of the banks corresponding to the intermediate nodes comprise the clearing currencies of the receiving banks in the cross-border remittance information, the cross-border clearing path knowledge graph is constructed, the cross-border clearing path is generated to meet the requirements of cross-border clearing in various scenes, the cross-border financial service level is improved, and the problems faced by current cross-border clearing are solved.

Description

Technical Field

[0001] This invention belongs to the field of financial technology, and more specifically, relates to a method, system, electronic device, and storage medium for recommending cross-border clearing paths. Background Technology

[0002] With the deepening of globalization, the scale of cross-border financial business continues to expand, and cross-border payment and clearing has become an important link connecting the global economy. However, the current cross-border clearing system faces many challenges and risks.

[0003] First, frequent changes in the international political and economic landscape have introduced uncertainty into cross-border transactions. Second, traditional cross-border clearing path selection is often based on fixed rules and the historical experience of sales personnel, making it difficult to adapt to the rapidly changing international environment. Furthermore, cross-border transactions involve numerous intermediate steps and complex processes, resulting in high transaction costs and low efficiency. Therefore, there is an urgent need for a cross-border clearing path recommendation method to improve the level of cross-border financial services and address the current challenges facing cross-border clearing. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a method, system, electronic device, and storage medium for recommending cross-border clearing paths. By constructing a cross-border clearing path knowledge graph, cross-border clearing paths are generated to meet the needs of cross-border clearing in various scenarios, improve the level of cross-border financial services, and solve the current problems faced by cross-border clearing.

[0005] The first aspect of this application discloses a method for recommending cross-border clearing routes, including:

[0006] Based on data from multiple data sources, a cross-border clearing path knowledge graph is constructed. The multiple data sources include at least SWIFTGPI, CIPS (Cross-border Interbank Payment System), and the banking system of CIPS. The cross-border clearing path knowledge graph includes at least the clearing path relationship, clearing fee relationship, and clearing timeliness relationship between bank nodes.

[0007] Based on the cross-border remittance information, the target cross-border clearing path is determined from the cross-border clearing path knowledge graph, and the target cross-border clearing path is visualized. The number of the target cross-border clearing paths is greater than or equal to 1. The head node of the target cross-border clearing path is the remitting bank of the cross-border remittance information, the tail node is the receiving bank of the cross-border remittance information, and the clearing currency of the banks corresponding to the intermediate nodes includes the clearing currency of the receiving bank.

[0008] Optionally, the above-mentioned cross-border clearing path recommendation method, after constructing a cross-border clearing path knowledge graph based on data from multiple data sources, also includes:

[0009] Determine the update granularity of the cross-border clearing path knowledge graph;

[0010] Based on the update granularity, the cross-border clearing path knowledge graph is updated using newly generated data from multiple data sources within the update granularity.

[0011] Optionally, in the above-mentioned cross-border clearing path recommendation method, constructing a cross-border clearing path knowledge graph based on data from multiple data sources includes:

[0012] Data from multiple data sources is cleaned to obtain cross-border clearing data;

[0013] The entities and relationships in the cross-border clearing data are extracted to obtain the cross-border clearing path knowledge graph. The entities include time entities, link entities, international bank identification code (BIC) entities, and composite BIC entities. The relationships include transaction statistics relationships, solid line flow relationships, dashed line flow relationships, and currency flow relationships.

[0014] Optionally, in the above-mentioned cross-border clearing path recommendation method, the data from multiple data sources is cleaned to obtain cross-border clearing data, including:

[0015] Identify erroneous data in the data from multiple said data sources;

[0016] Missing values ​​in the erroneous data are filled in, and outliers and duplicate values ​​in the erroneous data are deleted.

[0017] Optionally, in the above-mentioned cross-border clearing path recommendation method, determining the target cross-border clearing path from the cross-border clearing path knowledge graph based on cross-border remittance information includes:

[0018] Determine the query conditions corresponding to the cross-border remittance information, wherein the query conditions include at least: the remitting bank, the receiving bank, and the clearing currency of the cross-border remittance information;

[0019] The cross-border clearing paths in the cross-border clearing path knowledge graph that meet the query conditions are taken as the target cross-border clearing paths.

[0020] Optionally, in the above-mentioned cross-border clearing path recommendation method, visualizing the target cross-border clearing path includes:

[0021] The target visualization display type of the target cross-border clearing path is determined. The target visualization display type is one of the following: commonly used clearing path recommendation display, all clearing path recommendation display, shortest clearing path recommendation display, least cost clearing path recommendation display, shortest time clearing path display, and risk level clearing path recommendation display.

[0022] The target cross-border clearing path is visualized according to the target visualization display type.

[0023] The second aspect of this application discloses a cross-border clearing route recommendation system, comprising:

[0024] The construction module is used to build a cross-border clearing path knowledge graph based on data from multiple data sources, including at least SWIFT GPI, the RMB Cross-border Payment System CIPS, and banking systems. The cross-border clearing path knowledge graph includes clearing path relationships, clearing fee relationships, and clearing timeliness relationships between bank nodes.

[0025] The display module is used to determine the target cross-border clearing path from the cross-border clearing path knowledge graph based on the cross-border remittance information, and to visualize the target cross-border clearing path. The number of the target cross-border clearing paths is greater than or equal to 1. The head node of the target cross-border clearing path is the remitting bank of the cross-border remittance information, the tail node is the receiving bank of the cross-border remittance information, and the clearing currency of the banks corresponding to the intermediate nodes includes the clearing currency of the receiving bank.

[0026] Optionally, the aforementioned cross-border clearing path recommendation system may also include:

[0027] The determination module is used to determine the update granularity of the cross-border clearing path knowledge graph;

[0028] An update module is used to update the cross-border clearing path knowledge graph based on the update granularity, using newly generated data from multiple data sources within the update granularity.

[0029] A third aspect of this application discloses an electronic device, comprising:

[0030] One or more processors;

[0031] A storage device on which one or more programs are stored;

[0032] When one or more programs are executed by one or more processors, the one or more processors implement the cross-border clearing path recommendation method as described in any one of the first aspects of this application.

[0033] The fourth aspect of this application discloses a storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the cross-border clearing path recommendation method as described in any one of the first aspects of this application.

[0034] As can be seen from the above technical solution, the cross-border clearing path recommendation method provided by the present invention includes: constructing a cross-border clearing path knowledge graph based on data from multiple data sources, including at least SWIFT GPI, the RMB Cross-border Payment System (CIPS), and banking systems; the cross-border clearing path knowledge graph includes clearing path relationships, clearing fee relationships, and clearing timeliness relationships between bank nodes; determining a target cross-border clearing path from the cross-border clearing path knowledge graph based on cross-border remittance information, and visually displaying the target cross-border clearing path; the number of target cross-border clearing paths is greater than or equal to 1; the head node of the target cross-border clearing path is the remitting bank of the cross-border remittance information, the tail node is the receiving bank of the cross-border remittance information, and the clearing currency of the banks corresponding to the intermediate nodes includes the clearing currency of the receiving bank; by constructing a cross-border clearing path knowledge graph, cross-border clearing paths are generated to meet the needs of cross-border clearing in various scenarios, improving the level of cross-border financial services, and thus solving the current problems faced by cross-border clearing. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 This is a flowchart of a cross-border clearing path recommendation method provided in an embodiment of the present invention;

[0037] Figure 2 This is a flowchart illustrating the construction of a cross-border clearing path knowledge graph provided in an embodiment of the present invention;

[0038] Figure 3 This is a flowchart illustrating the determination of a target cross-border clearing path provided by an embodiment of the present invention;

[0039] Figure 4 This is a visual flowchart illustrating a target cross-border clearing path provided by an embodiment of the present invention;

[0040] Figure 5 This is a flowchart of another cross-border clearing path recommendation method provided by an embodiment of the present invention;

[0041] Figure 6 This is a cross-border clearing path knowledge graph network diagram provided in an embodiment of the present invention;

[0042] Figure 7 This is a schematic diagram illustrating the construction process of a cross-border clearing path knowledge graph provided in an embodiment of the present invention;

[0043] Figure 8 This is a schematic diagram illustrating the construction process of another cross-border clearing path knowledge graph provided in an embodiment of the present invention;

[0044] Figure 9 This is a schematic diagram of the source channel and destination channel in a transaction channel provided by an embodiment of the present invention;

[0045] Figure 10 This is a schematic diagram of a cross-border clearing path recommendation system provided in an embodiment of the present invention;

[0046] Figure 11 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] In this application, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0049] First, it should be noted that SWIFT (Society for Worldwide Interbank Financial Telecommunications) is an international interbank cooperative organization that aims to promote the standardization and automation of global financial communications. It has developed communication standards between financial institutions.

[0050] SWIFT GPI (Global Payment Innovation) is a new global payment standard developed by the SWIFT organization. Each GPI payment message carries a unique end-to-end identifier, and the payment path is tracked through the GPI Tracker. The payment status is reported as needed, and it features priority processing, tracking feedback, transparent fees, and complete information.

[0051] CIPS (Cross-border Interbank Payment System) is a wholesale payment system approved by the People's Bank of China, dedicated to RMB cross-border payment and clearing. CIPS participants are divided into direct participants and indirect participants, each assigned a system bank code as their unique identifier within the CIPS system. Direct participants open accounts with CIPS and send and receive transactions directly through CIPS, while indirect participants obtain services from CIPS through direct participants.

[0052] The Business Identifier Code (BIC) is an international standard developed by SWIFT for identifying financial institutions. Each BIC consists of 8 or 11 characters, encoded as follows: 4-digit bank code + 2-digit country code + 2-digit city code + 3-digit branch code. For example, ABOCNCNBJ190 represents the Guangdong Branch of the Agricultural Bank of China. Typically, a branch code of XXX indicates the headquarters of the financial institution, such as ABOCCNBJXXX representing the head office of the Agricultural Bank of China. Using the BIC, the sending bank and receiving bank can be uniquely identified. If the BIC is incorrect or does not exist, payment cannot be made via bank transfer.

[0053] Cross-border remittances often rely on complex banking agent networks. Banks clear cross-border funds by opening agent accounts. The bank providing the agent account service is the account bank of the bank holding the agent account, while the bank holding the agent account is the opening bank of the bank providing the agent account service. For example, if Bank A opens an agent account with Bank B, and Bank B opens an agent account with Bank C, then for Bank A, Bank B is its account bank; for Bank B, Bank A is its opening bank; and for Bank C, Bank B is its opening bank.

[0054] In developing this application, the inventors discovered that traditional clearing path selection methods have the following limitations: First, they lack the ability to dynamically assess real-time risk factors; second, they struggle to handle massive amounts of transaction data and complex relationships; third, risk control is insufficient, as manual selection cannot comprehensively assess various potential dangers and effectively predict potential transaction risks; fourth, manual path selection is inefficient, requiring sales staff to manually filter through numerous clearing paths, which is time-consuming and prone to errors; and fifth, it is highly experience-dependent, making it difficult for novice sales staff to quickly master optimal path selection techniques. These problems seriously affect the security and efficiency of cross-border transactions.

[0055] To address this, embodiments of the present invention provide a method, system, electronic device, and storage medium for recommending cross-border clearing paths. By constructing a cross-border clearing path knowledge graph, cross-border clearing paths are generated to meet the needs of cross-border clearing in various scenarios, thereby improving the level of cross-border financial services, solving the current challenges faced by cross-border clearing, and enhancing the security and efficiency of cross-border transactions.

[0056] Please see Figure 1 The recommended method for cross-border clearing mainly includes the following steps:

[0057] S101. Construct a knowledge graph of cross-border clearing paths based on data from multiple data sources.

[0058] In constructing a knowledge graph of cross-border clearing paths, the selection of data sources is crucial, as different data sources directly affect the final analysis results. The multiple data sources in this application include at least SWIFT GPI, CIPS, and banking systems.

[0059] In constructing a knowledge graph of cross-border clearing paths, SWIFT GPI provides at least SWIFT GPI message data. SWIFT GPI message data is a standardized collection of information used for global payments, designed to improve payment efficiency and transparency. It uses SWIFT's standard message formats (such as MT103 and MT202) and includes information such as amount, currency, remitter, recipient, and bank, enabling real-time tracking and recording of payment paths.

[0060] In constructing the knowledge graph of cross-border clearing paths, CIPS provides at least CIPS message data and CIPS participant information data. CIPS message data is key information transmitted in the Cross-border Interbank Payment System (CIPS) for processing cross-border RMB transactions. Unlike SWIFT GPI message data, it specifically handles RMB transactions and includes information such as amount, currency, remitter, payee, and bank. It offers advantages such as multiple payment methods, efficient clearing, and security. CIPS participant information data includes direct and indirect participants. Direct participants can be the banks corresponding to the head and tail nodes of the clearing path, while indirect participants can be the banks corresponding to the intermediate nodes of the clearing path.

[0061] In the process of constructing a knowledge graph of cross-border clearing paths, the banking system should provide at least the following data: cross-border remittance transaction data, international bank identification code data, sanctions screening transaction hit data, refund transaction data, account bank data, and account opening bank data.

[0062] The cross-border clearing path knowledge graph includes clearing path relationships, clearing fee relationships, and clearing timeliness relationships between bank nodes.

[0063] In practical applications, the cross-border clearing path knowledge graph is a network structure composed of BICs, where each point in the network represents a bank (or financial institution entity).

[0064] In some embodiments, the specific process of step S101, constructing a cross-border clearing path knowledge graph based on data from multiple data sources, is as follows: Figure 2 As shown, it mainly includes steps S201 and S202:

[0065] S201. Clean the data from multiple data sources to obtain cross-border clearing data.

[0066] Cross-border clearing data should include at least the relevant elements of cross-border clearing, such as clearing path information, transaction information, exchange rate information, fee information, and legal and regulatory information.

[0067] In practical applications, data can be collected from various data sources separately, resulting in data from multiple sources. Since each data source has a different data format, constituting heterogeneous data, it needs to be processed into cross-border clearing data with the same structure before subsequent training. Therefore, it is necessary to clean the data from multiple data sources to remove noise, errors, and redundant information, thereby improving data quality.

[0068] In some embodiments, the specific process of step S201, cleaning data from multiple data sources to obtain cross-border clearing data, mainly includes steps S301 and S302:

[0069] S301. Identify erroneous data in data from multiple data sources.

[0070] Erroneous data includes missing values, outliers, and duplicate values. Some erroneous data (such as outliers) occurs due to historical reasons; this type of erroneous data does not help with model training and therefore needs to be removed. Other erroneous data (such as missing values) are simply missing values, such as empty words or empty fields.

[0071] In practical applications, data from multiple data sources can be identified to determine unexpected extreme values ​​or abnormal data, duplicate data, and data with incorrect formatting, thus obtaining erroneous data.

[0072] S302. Fill in missing values ​​in the erroneous data and delete outliers and duplicate values ​​in the erroneous data.

[0073] In practical applications, the mean, median, and mode of the data source can be used to fill in missing values ​​and delete outliers and duplicate values.

[0074] The above processing enables high-quality processing of data collected from multiple data sources, laying a solid foundation for subsequent data analysis and applications.

[0075] It is understandable that data cleaning, standardization, and deduplication can integrate multi-source data into a unified cross-border clearing data warehouse.

[0076] It should be noted that, since the data from each data source is updated regularly, it is also necessary to periodically obtain data from multiple data sources and perform the aforementioned cleaning to obtain cross-border clearing data.

[0077] S202. Extract entities and relationships from cross-border clearing data to obtain a cross-border clearing path knowledge graph.

[0078] Entities include time entities, link entities, BIC entities, and composite BIC entities. Relationships include transaction statistics relationships, solid line flow relationships, dashed line flow relationships, and currency flow relationships. Among them, BIC entities represent banks with clearing transactions.

[0079] Based on cross-border clearing data, entities and relationships are extracted from the cross-border clearing data to construct a cross-border clearing knowledge graph.

[0080] In practical applications, knowledge graphs describe concepts and their relationships in the objective world in a structured form. Using a graph as the carrier, nodes represent "entities" and edges represent "relationships." Both entities and relationships can have their own attributes. The basic building block of a knowledge graph is the "entity-relationship-entity" triple. Entities are interconnected through relationships, forming a network-like knowledge structure.

[0081] The cross-border clearing knowledge graph in this application is a knowledge graph with at least cross-border clearing banks as nodes, including clearing path relationships, clearing fee relationships, and clearing timeliness relationships between various banks. The cross-border clearing knowledge graph includes several nodes and edges. Nodes represent entities in the clearing path (such as cross-border clearing banks), and edges represent relationships between entities (such as transaction relationships between banks). The cross-border clearing knowledge graph can intuitively display the complex relationships and potential value of cross-border clearing paths.

[0082] It should be noted that this application can utilize artificial intelligence technology to construct a knowledge graph of cross-border clearing paths, enabling in-depth analysis of cross-border transaction data and identification of potential relationships.

[0083] S102. Determine the target cross-border clearing path from the cross-border clearing path knowledge graph based on cross-border remittance information, and visualize the target cross-border clearing path.

[0084] The number of target cross-border clearing paths is greater than or equal to 1. The head node of the target cross-border clearing path is the remitting bank of the cross-border remittance information, the tail node is the receiving bank of the cross-border remittance information, and the clearing currency of the banks corresponding to the intermediate nodes includes the clearing currency of the receiving bank.

[0085] In some embodiments, the specific process of determining the target cross-border clearing path from the cross-border clearing path knowledge graph based on cross-border remittance information in step S102 is as follows: Figure 3 As shown, it mainly includes steps S401 and S402:

[0086] S401. Determine the query conditions corresponding to cross-border remittance information.

[0087] The search criteria should include at least the remitting bank, receiving bank, and clearing currency of cross-border remittance information.

[0088] In practical applications, cross-border remittance information includes the remitting bank, the receiving bank, the remittance amount, and the settlement currency.

[0089] S402. Select the cross-border clearing paths in the cross-border clearing path knowledge graph that meet the query conditions as the target cross-border clearing paths.

[0090] In practical applications, the query conditions should include at least the remitting bank, receiving bank, and clearing currency of the cross-border remittance information. The query should be performed on the cross-border clearing path knowledge graph, and the cross-border clearing path that meets the query conditions in the cross-border clearing path knowledge graph should be used as the target cross-border clearing path.

[0091] It is understandable that a cross-border clearing path knowledge graph is used to generate a target cross-border clearing path that meets the requirements through graph algorithms.

[0092] It should be noted that in practice, query conditions can also include specified intermediate nodes to meet the needs of different application scenarios.

[0093] In some embodiments, the specific process of visualizing the target cross-border clearing path in step S102 is as follows: Figure 4 As shown, it mainly includes steps S501 and S502:

[0094] S501, Target visualization display type for determining the target cross-border clearing path.

[0095] The target visualization display type is one of the following: commonly used liquidation path recommendation display, all liquidation path recommendation display, shortest liquidation path recommendation display, least cost liquidation path recommendation display, shortest time liquidation path display, and risk level liquidation path recommendation display.

[0096] In practical applications, the type of target visualization can be determined based on the specific application scenario requirements, and can be any of the above.

[0097] Cross-border transactions often involve many risks. For both customers and banks, ensuring the safety of funds is always the most important thing. Therefore, the familiar and frequently used clearing path is often the first choice. In this case, the commonly used clearing paths can be recommended and displayed as the target visualization type. The target cross-border clearing paths will be sorted from high to low according to the number of transactions along the clearing path.

[0098] To enable sales staff to anticipate the potential trajectory of a particular transaction and be well-informed, a comprehensive display of all clearing paths can be presented as a target visualization, showcasing all target cross-border clearing paths.

[0099] Considering that each intermediary bank carries the risk of customers incurring additional fees, it is advisable to select the shortest clearing path from the outset to minimize intermediary fees and transaction time for customers. This can be achieved by displaying the shortest clearing path as the target visualization type, with the target cross-border clearing paths sorted from lowest to highest number of intermediary transactions.

[0100] Cross-border remittances often pass through multiple intermediary banks before reaching the receiving bank. Currently, the fee rules for these intermediary banks are not transparent, making it impossible to predict transfer fees. Customers only know the fees deducted after receiving the remittance. However, this application can extract interbank fee deductions from SWIFT GPI message data. By fitting these interbank fee deductions, it can obtain fee estimates for different clearing paths. If customers want the clearing path with the lowest transaction fees, they can display a recommendation of the lowest-cost clearing path as the target visualization type. In this case, target cross-border clearing paths are sorted from lowest to highest transaction fees.

[0101] Compared to domestic remittances, cross-border remittances have a longer processing time, ranging from a few minutes to several days. The arrival time of a remittance depends on the processing speed of the intermediary bank. If customers want the shortest remittance time, they can display the shortest clearing path as the target visualization type. In this case, the target cross-border clearing paths are sorted from shortest to longest clearing time.

[0102] Because this application incorporates data on sanction screening transactions and refund transactions, it allows for risk labeling, sanction node labeling, and refund transaction labeling of clearing paths. If clients want to understand whether a clearing path faces sanction or refund risks, they can identify these risks and proactively mitigate them. The risk level of the clearing path can be recommended and displayed as a target visualization type, sorted from highest to lowest based on risk labeling, sanction node labeling, and refund transaction labeling. Additionally, preset thresholds can be set to alert clients to clearing paths with a certain percentage of refunds, thus mitigating refund risks.

[0103] S502. Visualize the target cross-border clearing path according to the target visualization display type.

[0104] In practical applications, once the target cross-border clearing path is determined, the target cross-border clearing path can be visualized according to the target visualization type to meet the recommendation needs of different scenarios.

[0105] It should be noted that general graph database visualization tools can be used to visualize the cross-border clearing path knowledge graph. In the interactive analysis application of the cross-border clearing path knowledge graph, the graph engine can be used to generate clearing paths to meet the clearing path recommendation function in various scenarios.

[0106] It should also be noted that the main purpose of constructing the cross-border clearing path knowledge graph is for querying. Therefore, this application can use a database plugin, a no-code, keyword-based graph data visualization tool, to visualize the relationships in the cross-border clearing path knowledge graph.

[0107] It should be noted that since all cross-border transaction information that has occurred in this application, including currency, amount, number of transactions, etc., is statistically summarized into the link entity, BIC entity - solid line flow - BIC entity relationship, it is possible to view all related queries. For example, you can find statistics on the most frequently used clearing paths for cross-border remittances from a commercial bank to a certain country, which clearing path is most frequently used for the largest and smallest amount of transactions, and which BIC entity is a key node on the clearing path, etc.

[0108] Based on the above principles, the cross-border clearing path recommendation method provided in this embodiment includes: constructing a cross-border clearing path knowledge graph based on data from multiple data sources, including at least SWIFT GPI, CIPS, and banking systems; the cross-border clearing path knowledge graph includes clearing path relationships, clearing fee relationships, and clearing timeliness relationships between bank nodes; determining a target cross-border clearing path from the cross-border clearing path knowledge graph based on cross-border remittance information, and visually displaying the target cross-border clearing path; the number of target cross-border clearing paths is greater than or equal to 1; the head node of the target cross-border clearing path is the remitting bank of the cross-border remittance information, the tail node is the receiving bank of the cross-border remittance information, and the clearing currency of the banks corresponding to the intermediate nodes includes the clearing currency of the receiving bank; by constructing a cross-border clearing path knowledge graph, cross-border clearing paths are generated to meet the needs of cross-border clearing in various scenarios, improving the level of cross-border financial services, thereby solving the current difficulties faced by cross-border clearing, and improving the security and efficiency of cross-border transactions.

[0109] Furthermore, by constructing a global cross-border clearing path knowledge graph, combined with real-time data and smart contracts, dynamic adjustments and optimal matching of paths can be achieved, further enhancing the transparency and efficiency of cross-border transactions. This will not only help enterprises gain a more advantageous position in global competition but will also drive the digital transformation of the entire financial industry, providing more convenient, secure, and efficient services for participants in cross-border transactions.

[0110] It's worth noting that cross-border transactions are unique in that a single transaction often involves two or more commercial banks. Each bank can only choose to transfer funds to its next bank, but has no say in where the next bank should direct the funds. This leads to longer clearing paths and increased risks. Therefore, shortening clearing paths, when appropriate, should reduce these risks. In cross-border transactions, all banks involved can be considered nodes in a clearing network graph. By continuously exploring the potential value of existing SWIFT GPI and CIPS message data and improving the utilization rate of this clearing data, relevant bank nodes can be drawn in the clearing network graph, depicting the clearing relationships between banks. This is a natural graph structure, and graph databases excel in relation traversal and path search. Therefore, this application focuses on constructing a cross-border clearing path knowledge graph, using a graph engine to generate clearing paths to meet the clearing path recommendation needs in various scenarios. This application already supports multiple liquidation path recommendation scenarios, including commonly used liquidation path recommendation, all liquidation path recommendation, shortest liquidation path recommendation, shortest cost liquidation path recommendation, shortest time liquidation path recommendation, and risk level liquidation path identification. It also supports various liquidation data statistical query needs.

[0111] It is understood that the purpose of the cross-border clearing path recommendation method provided in this application is to: utilize artificial intelligence technology to construct an intelligent cross-border clearing path knowledge graph, achieve intelligent optimization of cross-border transaction paths, monitor transaction risks in real time, reduce transaction costs, and improve clearing efficiency. Specific objectives include:

[0112] 1. Enable in-depth analysis and correlation mining of cross-border transaction data to identify potential business connections.

[0113] 2. Establish a dynamic risk assessment model to monitor in real time the impact of changes in the international political and economic environment on cross-border transactions.

[0114] 3. Provide intelligent path recommendation service to optimize clearing path selection and reduce transaction costs.

[0115] 4. Enable early warning of transaction risks and improve the security of cross-border transactions.

[0116] 5. Provide decision support for financial institutions and corporate users, and improve the level of cross-border financial services.

[0117] By achieving the above objectives, this application will provide strong technical support for cross-border financial business, helping enterprises to avoid transaction risks, reduce operating costs, and enhance international competitiveness.

[0118] In practical applications, since data from different data sources changes in real time, if only the initially constructed cross-border clearing path knowledge graph is used as the basis for queries, the target cross-border clearing path obtained will become significantly inconsistent with the actual situation over time. Therefore, regular updates to the cross-border clearing path knowledge graph are particularly important. In this regard, this application provides another embodiment, such as... Figure 5 As shown, after performing step S101 and constructing a cross-border clearing path knowledge graph based on data from multiple data sources, steps S601 and S602 are also included:

[0119] S601. Determine the update granularity of the cross-border clearing path knowledge graph.

[0120] In practical applications, the update granularity of the cross-border clearing path knowledge graph represents the corresponding update cycle of the cross-border clearing path knowledge graph.

[0121] It should be noted that the update granularity of the cross-border clearing path knowledge graph can be determined based on the data update cycles of multiple data sources, or it can be preset. For example, the update granularity can be a calendar month; of course, it is not limited to this, and can also be determined according to the application environment and user needs, all of which are within the scope of protection of this application.

[0122] S602. Based on the update granularity, the knowledge graph of cross-border clearing paths is updated using newly generated data from multiple data sources within the update granularity.

[0123] In practical applications, the knowledge graph of cross-border clearing paths can be updated using newly generated data from multiple data sources at the update granularity, allowing the knowledge graph to continuously self-correct based on the accumulation of actual business data.

[0124] For example, this application introduces a time entity, which can support monthly updates of existing knowledge graph data. For transaction data in a new month, it can automatically add new time entities. For existing link entities, BIC entities, and composite BIC entities, new transaction statistics relationships will be established with the time entities. For non-existent link entities, BIC entities, and composite BIC entities, new entities will be created directly, and corresponding relationships will be established.

[0125] The cross-border clearing path recommendation method provided in this embodiment constructs a cross-border clearing path knowledge graph based on data from multiple data sources, and then determines the update granularity of the cross-border clearing path knowledge graph. Based on the update granularity, the cross-border clearing path knowledge graph is updated using newly generated data from multiple data sources within the update granularity. This ensures that the cross-border clearing path knowledge graph used for retrieval continuously self-corrects based on the accumulation of actual business data, guaranteeing the timeliness and effectiveness of the obtained target cross-border clearing path results.

[0126] To facilitate understanding, the construction process of a cross-border clearing knowledge graph is illustrated below with a specific example:

[0127] Step 1: Define the knowledge graph ontology model of cross-border clearing paths

[0128] The knowledge graph of cross-border clearing paths mainly consists of four types of entities: time entities, link entities, international bank identification code entities, and composite international bank identification code entities, and four types of relationships: transaction statistics relationships, solid line flow relationships, dashed line flow relationships, and currency flow relationships.

[0129] The time entity representation refers to the time unit used to summarize data. For example, data can be summarized at the monthly granularity, such as 202501 and 202502, which are time entities.

[0130] A link entity represents a complete clearing link. For example, bank1->bank2->bank3 is a link. It is considered a complete entity. This link also includes many attributes, such as link number, link content, countries passed through the link, head node of the link, tail node of the link, virtual and physical direction of the link, transaction currency of the link, length of the link, and whether the link is a risk link.

[0131] A BIC entity represents a bank entity extracted from a data chain. For example, bank1, bank2, and bank3 extracted from the chain bank1->bank2->bank3 all belong to BIC entities, representing an actual bank or other financial institution. They are all extracted from cross-border clearing data. A BIC entity includes elements such as BIC number, BIC content, city of origin, country of origin, full name of BIC, length of BIC, whether the BIC is an account bank, whether the BIC is the opening bank, and whether the BIC has been sanctioned.

[0132] A composite BIC entity represents a link composed of two international bank identification code entities. For example, the link bank1->bank2->bank3 can be further broken down into "smaller links" such as bank1->bank2 and bank2->bank3, from which a lot of information can be extracted. In addition to the elements that are consistent with international bank identification code entities, the composite international bank identification code entity also includes the length element of the link composed of two international bank identification code entities.

[0133] Different relationships exist between the various entities, including: transaction statistics relationships, solid line flow relationships, dashed line flow relationships, and currency flow relationships.

[0134] Transaction statistics relationship: This relationship connects a time entity on one end and a BIC entity, composite BIC entity, or link entity on the other. For example, combining... Figure 6 In the cross-border clearing chain knowledge graph, a time entity pointing to a BIC entity represents a transaction statistics relationship. This relationship includes elements such as the total number of transactions, the total transaction amount, and the currency of the transactions processed by that BIC entity in the clearing chain at that time. Time entities pointing to composite BIC entities and time entities pointing to chain entities have the same meaning as above; they all represent the total transaction information for a specific BIC entity, composite BIC entity, or chain within a given time period.

[0135] Solid line flow relationship: This relationship is between BIC entities, including the relationship between a BIC entity and a composite BIC entity. The relationship attributes include time, currency, total number of transactions, and total transaction amount, recording the transaction information flowing between a specific BIC entity and other BIC entities within a certain period. For example, in 202501, this shows the number of transactions and transaction amount for a specific currency between bank1 and bank2.

[0136] Dashed line flow relationship: Due to limitations in SWIFT GPI and CIPS message data, only partial transit information on the clearing path can be obtained. Therefore, some path selection may be ambiguous. To explore potential possibilities, this part of the data information can be retained. Here, the path that is certain to be able to transit and clear is defined as the realized flow relationship, and the path that is less certain to be able to transit and clear is defined as the dashed line flow relationship. The attributes of the dashed line flow relationship are consistent with those of the realized flow relationship.

[0137] Currency Flow Relationship: This relationship is between BIC entities. The attributes of the relationship include the currency being traded and the trading channel (SWIFT or CIPS). Through this relationship, the clearing flow of the transaction amount for a certain currency can be clearly seen.

[0138] Special Note: In addition to SWIFT GPI and CIPS message data, this application also incorporates data from an internal sanctions screening system of a bank. This data can detect whether a transaction link has been flagged by the screening system. If it has, it indicates that certain information in the link is at risk and therefore requires additional marking. Simultaneously, internal refund transaction data from the bank is also included. Refunds can be due to various reasons, such as improper operation by business personnel or systemic reasons. This information will be marked in the link entity, BIC entity, and in transaction statistics relationships, solid line flow relationships, and dashed line flow relationships.

[0139] The aforementioned entity and relationship definitions allow users to view clearing path networks for any currency and any BIC. Users can input the starting and ending BICs for cross-border clearing path recommendations, such as recommendations for frequently used clearing paths, lists of all clearing paths, and the shortest-fee clearing path. Furthermore, the introduction of time-based entities enables incremental updates to cross-border clearing graph relationships and automatic statistics of cross-month transaction relationships. Users can view transaction statistics for a specific link within a specified time period, as well as transaction statistics between specific BICs. Additionally, because the relationships include attributes such as whether they have been hit by the sanctions screening system or whether they have been refunded, this application also supports the identification of risky links.

[0140] Using data from a specific commercial bank as a perspective, this project aggregates cross-border clearing paths from multiple data sources at a monthly granularity, reconstructing the cross-border transaction relationship network of each bank within the current month. Specifically, banks with transaction clearing in the current month (BIC entities) are treated as knowledge graph entities, bank tags (such as bank name, country of origin, and whether the bank is an account bank) are treated as entity attributes, and the total amount and number of bank clearing transactions in the current month are treated as entity relationship attributes.

[0141] For example, a cross-border clearing path knowledge graph network diagram can be as follows: Figure 6 As shown.

[0142] The following simple example further illustrates the above. Consider the following data: Company A in country S needs to pay business fees to Company B in country Y. Company A remits funds to Company B through bank 1, and Company B receives the funds from bank 3. For Company A, the remittance originates from bank 1, making bank 1 the starting point of this cross-border clearing transaction. For Company B, the receipt of funds is from bank 3, making bank 3 the ending point of this cross-border clearing transaction. However, what Company A and Company B are unaware of is that their funds may actually be transferred through different clearing paths due to different processing methods by bank 1's staff. For example, the funds may need to pass through bank 2, because bank 1 may not have the authority to directly transfer funds to bank 3. There are several possible transfer options; for instance, bank 4 can also directly transfer funds to bank 3. Therefore, if bank 1's staff chooses to transfer funds to bank 4, then the funds will be transferred from bank 4 to bank 3.

[0143] The above examples, bank1->bank2->bank3 and bank1->bank4->bank3, represent two cross-border clearing paths. However, historical data reveals more complex scenarios. For instance, a staff member at bank1 might entrust funds to bank2 for transfer to bank3, but bank2 lacks the authority to directly transfer funds to bank3. Instead, bank2 transfers the funds to bank5, bank5 to bank6, and then bank6 delivers the funds to bank3. In this case, bank1->bank2->bank5->bank6->bank3 becomes a new clearing path. Such complex situations are numerous in historical data. This occurs because there are hundreds of thousands, even millions, of banks worldwide, facing numerous geopolitical conflicts and risks. Furthermore, banks often act as unconnected entities; bank1 is unaware of bank2's counterparties. Therefore, banks typically entrust funds to banks they trust (those with which they have accounts). However, this involves many choices, such as a bank having hundreds of account holders.

[0144] By using SWIFT GPI and CIPS message data, we can obtain part of the remittance clearing process. Therefore, we can identify that there is a transaction relationship between Bank A and Bank B. We can then infer that Bank A and Bank B can transfer funds, and thus we can transfer the funds to a bank that can deliver them to the final receiving bank.

[0145] Different banks have different clearing currencies; for example, some banks can clear the ruble, while others cannot. Therefore, understanding which bank can clear which currency is crucial. Historical data shows that in a certain month (e.g., January 2025), there were 10 remittances cleared via bank1->bank2->bank3, primarily in USD and CNY, and 14 remittances cleared via bank1->bank4->bank3, primarily in USD and EUR. Therefore, it can be inferred that if USD is remitted from bank1 to bank3, bank1 staff can choose to have the funds processed by either bank2 or bank4; conversely, if EUR is remitted, bank1 staff can choose to have the funds processed by bank4.

[0146] By analyzing historical data, we can calculate the total number of clearing transactions and their amounts for each clearing path in a given month, the number of banks involved in that path, the countries these banks belong to, the currencies used, the transaction fees, and the total transit time. All this information is stored in the cross-border clearing path knowledge graph in the form of entities and relationships.

[0147] Step Two: Processing Knowledge Graph Data on Cross-Border Clearing Paths

[0148] Since the data sources of the cross-border clearing path knowledge graph are multiple data sources, a distributed relational database can be used to store multi-source data, classify it according to different path sources, integrate the paths with batch programs, extract entities and relationships, and complete the entity and relationship data processing of the cross-border clearing path knowledge graph.

[0149] Combination Figure 6 , Figure 7 and Figure 8 The main processes for processing cross-border clearing path knowledge graph data are as follows:

[0150] (1) Acquire knowledge sources and sort out the liquidation path

[0151] The knowledge sources for constructing the cross-border clearing path knowledge graph in this application mainly come from cross-border remittance transaction data, SWIFT GPI message data, CIPS message data, etc. Extracting effective cross-border clearing paths from multi-source and complex data is the primary challenge addressed in this application.

[0152] The cross-border clearing path knowledge graph is a network structure composed of BICs (Bill of Integrities). Each point in the network represents a specific commercial bank or financial institution entity. Therefore, from the perspective of this commercial bank or financial institution entity, the network structure is centered on that entity. This application takes the perspective of a specific commercial bank, assuming that the commercial bank is the "center" of the cross-border clearing path network structure. Therefore, cross-border clearing transactions involving this commercial bank can be divided into two types: "coming" and "going." Specifically, a clearing fund "comes" to this commercial bank for clearing, and a clearing fund "goes" from this commercial bank to the next bank or financial institution for clearing.

[0153] In cross-border clearing transactions, the channel through which clearing funds "come" to a commercial bank can be simply categorized as the SWIFT GPI channel, the CIPS channel, or other channels (referring to the commercial bank's internal business processing system). Similarly, the channel through which clearing funds "leave" a commercial bank can also be simply categorized as the SWIFT GPI channel, CIPS channel, or other channels. In analyzing cross-border clearing paths, only by assembling the "coming" and "going" paths can a clearing path from the perspective of a particular commercial bank be identified. Only by correctly identifying the clearing path can a solid foundation be laid for subsequent data processing, and a correct cross-border clearing path knowledge graph be constructed.

[0154] An example diagram illustrating the source and destination channels in a transaction is shown below. Figure 9 As shown.

[0155] Because there are many types of SWIFT GPI messages and CIPS messages, and different messages are processed in different ways, they need to be analyzed on a case-by-case basis. Therefore, this application only considers the MT103 and MT202 messages in the SWIFT GPI message data, and only considers customer remittance and institutional remittance messages in the CIPS message data.

[0156] For example, the liquidation path outlined based on business knowledge is shown in Table 1-1:

[0157] Table 1-1 Liquidation Path Summary Table

[0158] Source Channel -> Destination Channel Source Channel -> Destination Channel SWIFT(MT103)->SWIFT(MT103) CIPS (Customer Remittance) -> SWIFT (MT103) SWIFT (MT103) -> SWIFT (MT202) CIPS (Customer Remittance) -> SWIFT (MT202) SWIFT (MT202) -> SWIFT (MT103) CIPS (Institutional Remittances) -> SWIFT (MT103) SWIFT(MT202)->SWIFT(MT202) CIPS (Institutional Remittance) -> SWIFT (MT202) SWIFT (MT103) -> CIPS (Customer Remittance) CIPS (Customer Remittance) -> CIPS (Customer Remittance) SWIFT (MT103) -> CIPS (Institutional Remittance) CIPS (Customer Remittances) -> CIPS (Institutional Remittances) SWIFT (MT202) -> CIPS (Customer Remittance) CIPS (Institutional Remittance) -> CIPS (Customer Remittance) SWIFT (MT202) -> CIPS (Institutional Remittance) CIPS (Institutional Remittance) -> CIPS (Institutional Remittance) SWIFT (MT103) -> Other CIPS (Customer Remittance) -> Other SWIFT (MT202) -> Other CIPS (Institutional Remittances) -> Other Other -> SWIFT (MT103) Other -> CIPS (Customer Remittance) Other -> SWIFT (MT202) Other -> CIPS (Institutional Remittances) Other -> Other

[0159] In this application, cross-border remittance transaction data is the primary focus. Combined with SWIFT GPI message data and CIPS message data, the clearing paths are assembled and categorized according to the aforementioned source and destination channels. Taking a domestic commercial bank C as the "center" perspective, the clearing path of a genuine cross-border remittance transaction can be divided into the following three different types:

[0160] ① The clearing path ending at Commercial Bank C: Remitting Bank -> ... -> Indirect Paying Participating Bank -> Direct Paying Participating Bank -> Agent Bank 1 -> Agent Bank 2 -> Agent Bank 3 -> Direct Receiving Participating Bank -> Commercial Bank C

[0161] ② The clearing path starting from Commercial Bank C: Commercial Bank C -> Direct Paying Bank -> Agent Bank 1 -> Agent Bank 2 -> Agent Bank 3 -> Direct Receiving Bank -> ... -> Intermediate Bank -> Receiving Bank -> Lender's Account Opening Bank

[0162] ③ The clearing path of the transaction through Commercial Bank C: Remitting Bank -> ... -> Direct Participating Bank for Payment -> Agent Bank 1 -> Agent Bank 2 -> Agent Bank 3 -> Commercial Bank C -> Agent Bank 1 -> Agent Bank 2 -> Agent Bank 3 -> Direct Participating Bank for Receiving Payment ... -> Intermediate Bank -> Receiving Bank -> Lender's Account Opening Bank

[0163] Based on the above classification of clearing paths, the "->" symbol is used to represent direct clearing between commercial banks, while "...->" is used to represent clearing where it is unclear whether other banks will participate. In the relationship, these correspond to solid and dashed flow relationships between BIC entities, respectively. The relationship of whether direct clearing is possible is also marked in the link entity's link solid / dashed flow attribute.

[0164] (2) Entity extraction and relation extraction to acquire knowledge

[0165] This application has defined four entity types: time entity, link entity, BIC entity, and composite BIC entity; and four relationships: transaction statistics relationship, solid flow direction relationship, dashed flow direction relationship, and currency flow direction relationship. These four entity types and four relationships form eight types of knowledge triples, including time-transaction statistics-link, time-transaction statistics-BIC, time-transaction statistics-composite BIC, BIC-solid flow direction-BIC, BIC-dashed flow direction-BIC, BIC-currency flow direction-BIC, BIC-solid flow direction-composite BIC, and BIC-dashed flow direction-composite BIC, as shown in Table 1-2 below:

[0166] Table 1-2 Knowledge Triple Table Composed of Entities and Relationships

[0167] Entity class h - Entity class t Relationship type Entity class h - Entity class t Relationship type Time-link Transaction statistics BIC-BIC Dashed line flow direction Time - BIC Transaction statistics BIC-BIC Currency flow Time-composite BIC Transaction statistics BIC-Composite BIC solid line flow direction BIC-BIC solid line flow direction BIC-Composite BIC Dashed line flow direction

[0168] Below are the specific methods for entity extraction and relation extraction:

[0169] The time involved in the transaction will be extracted on a monthly basis. For example, 202501 and 202502 are time entities. Since the transaction statistics relationship includes the attribute of time, this application can view the specific transaction statistics relationship of a certain month at any time.

[0170] For each link identified from cross-border clearing transaction data, information such as platform number, source data table, link type, currency, amount, and whether it is a complete link is saved to ensure that the data source can be traced upstream. At the same time, identical links are clustered and merged, and then divided by time to obtain the actual number of transactions and transaction amount for each clearing path in each month, that is, the number of transactions and total transaction amount attributes in the transaction statistics relationship.

[0171] Cross-border clearing paths are composed of connections between BICs (Bill of Transactions). Therefore, all BICs can be extracted from the cross-border clearing path and then divided by time to obtain the actual number of transactions and transaction amount for each BIC each month. Furthermore, for subsequent clearing path recommendation functionality, adjacent BICs in the path can be merged into composite BICs, then divided by time, and the actual number of transactions and transaction amount for each month on this composite BIC can be calculated.

[0172] The link entity stores the virtual and real flow attributes of BICs in the link. Therefore, if BICs can be directly cleared, this application will define a solid flow relationship between the two BICs. If they cannot be directly cleared, this application will define a dashed flow relationship between the two BICs. The same applies to the solid and dashed flow relationships between BICs and composite BICs.

[0173] Since cross-border clearing paths support multiple currencies, this application also extracts the currency flow relationship between BICs from the actual link to support viewing the clearing network diagram of any currency.

[0174] (3) Supplementing and integrating knowledge

[0175] The data sources for the above entity and relation extraction are SWIFT GPI message data, CIPS message data, and cross-border remittance transaction data. However, relying solely on this data is insufficient to achieve the functionality intended by this application. The application also incorporates BIC data, account bank and opening bank data, CIPS participant information data, refund transaction data, transaction data detected by a bank's internal sanctions screening system, and data from officially released sanctioned banks and other financial institutions collected from the internet.

[0176] Because the BIC encoding specifies that a BIC consists of 11 digits, with digits 5-6 representing the country code and digits 7-8 representing the city code, the country and city of the BIC can be obtained from the BIC encoding. This application also introduces supplementary BIC data, including institution name information, account bank, and opening bank data to supplement the account bank and opening bank attribute information of the commercial bank's BIC. Officially released sanctioned BIC data collected from the internet can be used to mark whether a BIC is sanctioned, thus preventing subsequent fund flows to this BIC for liquidation. Since this application also includes composite BICs, composed of multiple BICs, which can be considered a small "link," a length attribute is added to the BIC's attributes to indicate whether it is a composite BIC. Processing CIPS participant information data can supplement missing BIC entities and also supplement missing BIC-currency flow-BIC relationships.

[0177] Refund transaction data typically includes a column recording the reason for the refund. Analysis of existing refund transactions reveals two main categories: the first is primarily related to institutional reasons, such as refunds processed according to internal regulations; the second is mainly related to operational errors, such as incorrect receiving accounts or receiving banks other than our bank's. By using keyword matching to analyze refund transactions, transactions with past refunds can be marked as either refunds due to institutional reasons or operational errors. The corresponding transaction chain is also marked, and the number and percentage of refund transactions due to institutional reasons or operational errors within each timeframe for that chain are statistically analyzed.

[0178] By introducing a sanctions screening system to detect transactions, transactions can be marked as risky transactions. At the same time, the links identified by these transactions can be marked as risky links. Furthermore, the number and percentage of transactions detected by the sanctions screening system can be calculated for each time period within these links.

[0179] (4) Update knowledge

[0180] Combination Figure 8 Since this application introduces time entities, it can support monthly updates of existing knowledge graph data. For transaction data in a new month, it can automatically add new time entities. For existing link entities, BIC entities, and composite BIC entities, new transaction statistics relationships will be established with time entities. For non-existent link entities, BIC entities, and composite BIC entities, new entities will be created directly, and corresponding relationships will be established.

[0181] Step 3: Data storage of cross-border clearing path knowledge graph

[0182] Entities and relationships are stored using a real-time graph database based on a traversal algorithm. Initial database data is constructed through a full data import module, and incremental database data is imported through an incremental data import module.

[0183] The data storage structure and query methods of graph databases are based on graph theory. The basic elements of a graph in graph theory are nodes and edges, which correspond to nodes and relations in graph databases, respectively, representing entities and relations in a knowledge graph. Graph data models can be divided into attribute graphs, hypergraphs, and triples, with attribute graphs being the most widely used. Real-time graph databases based on traversal algorithms employ the attribute graph model, which includes the following characteristics: it can store nodes and relations; both nodes and relations can have their own attributes, such as the BIC nodes in this application containing attributes like country and city, and transaction statistics relations containing attributes like transaction amount and number of transactions; relations are directional, always pointing from the start node to the end node. Graph databases offer the best performance in relation traversal and path search query applications because relations are "read" rather than calculated. In this application, the greatest application of the cross-border clearing path knowledge graph is in clearing path recommendation. In a pre-generated graph network, a start node and an end node are specified to find the most suitable clearing path for recommendation. Based on these characteristics, real-time graph databases based on traversal algorithms are suitable for storing cross-border clearing path knowledge graphs.

[0184] The storage process of the cross-border clearing path knowledge graph is also the process of importing data from the graph database. Importing data from the graph database mainly includes: directly using the Create statement of the graph database language Cypher (advantage: real-time data insertion; disadvantage: slow speed, requires writing Cypher statements); using Load CSV (advantage: direct import of CSV files, supports real-time insertion; disadvantage: slow import speed); using Batch Import (advantage: fast import speed, low resource consumption; disadvantage: requires stopping the graph database for offline import); and using full data import (advantage: very fast import speed; disadvantage: requires stopping the graph database for offline import). Since the entity and relation data involved in this application are in the millions, and there are requirements for import speed, this application uses the full data import method for the initial data. Subsequent monthly incremental updates of entities and relations are processed using the Load CSV method for real-time import, thus avoiding the need to stop the database.

[0185] In the cross-border clearing path knowledge graph data processing in step two above, the present invention will ultimately obtain the following data: entities are "time", "link", "BIC" and "composite BIC", and relationships are "transaction statistics relationship", "realization flow relationship", "dashed line flow relationship" and "currency flow relationship". All of these data are processed according to the data format required by the full import method, so they can be directly imported to create the initial database in one go.

[0186] In summary, this application integrates multi-source heterogeneous data and relies on big data computing capabilities and a graph database platform to complete the construction, processing, storage, and querying of a cross-border clearing path graph model with tens of thousands of entities and millions of relationships. It can meet the needs of querying and displaying the graph relationships of specified entities and targeted clearing paths, support users in mining potential value information in clearing paths, and support the generation of clearing paths to meet the needs of clearing path recommendations in various scenarios.

[0187] It's worth noting that the closest existing technology to solving the problem of cross-border clearing path recommendation is the invention patent titled "Method, System, Medium, and Device for Generating Cross-border Remittance Paths Based on SWIFT GPI Data." The main method involves using SWIFT GPI message data to obtain a point-to-point backtracking dataset containing backtracking directions and path values, storing it in a relational database, and having the front end input the receiving bank, currency, and recommendation type. The back end then generates the cross-border clearing path based on the input data and the backtracking dataset. However, this approach has the following drawbacks: 1. Limited data sources: Clearing data is primarily sourced from SWIFT GPI messages, while cross-border clearing data also includes CIPS messages, etc. The lack of training data reduces the number of recommended clearing paths, thus failing to achieve the desired optimal clearing path recommendation. 2. Low query efficiency for clearing path recommendations: Cross-border clearing paths are composed of multiple point-to-point connections. Querying the shortest path between the origin and destination using a relational database is less efficient than querying a real-time graph database based on traversal algorithms. 3. Low utilization rate of clearing data: The way clearing data is processed may lead to the loss of potentially valuable information on the clearing path, such as the transaction amount, number of transactions, and average processing time on the clearing path. 4. Clearing paths cannot be kept up-to-date: Due to changes in the external environment, banks on the clearing path are at risk of being sanctioned at any time, causing the original clearing path to become unusable. The recommended clearing path cannot be guaranteed to be up-to-date and valid, so the clearing path recommendation may not be accurate. The cross-border clearing path recommendation solution provided in the application can avoid the above shortcomings.

[0188] Alternatively, another embodiment of this application provides a cross-border clearing path recommendation system.

[0189] See Figure 10 The cross-border clearing path recommendation system includes:

[0190] The construction module 100 is used to construct a cross-border clearing path knowledge graph based on data from multiple data sources, including at least SWIFT GPI, CIPS, and banking systems. The cross-border clearing path knowledge graph includes clearing path relationships, clearing fee relationships, and clearing timeliness relationships between bank nodes.

[0191] The display module 200 is used to determine the target cross-border clearing path from the cross-border clearing path knowledge graph based on the cross-border remittance information, and to visualize the target cross-border clearing path. The number of the target cross-border clearing paths is greater than or equal to 1. The head node of the target cross-border clearing path is the remitting bank, the tail node is the receiving bank in the cross-border remittance information, and the clearing currency of the banks corresponding to the intermediate nodes includes the clearing currency of the receiving bank in the cross-border remittance information.

[0192] Optionally, the cross-border clearing route recommendation system may also include:

[0193] The determination module is used to determine the update granularity of the cross-border clearing path knowledge graph;

[0194] An update module is used to update the cross-border clearing path knowledge graph based on the update granularity, using newly generated data from multiple data sources within the update granularity.

[0195] For details on the specific working process and principles of each of the above modules, please refer to the cross-border clearing path recommendations provided in the above embodiments. They will not be elaborated here, and can be determined according to the actual situation, all of which are within the protection scope of this application.

[0196] In this embodiment, the construction module 100 is used to construct a cross-border clearing path knowledge graph based on data from multiple data sources, including at least SWIFT GPI, CIPS, and banking systems. The cross-border clearing path knowledge graph includes clearing path relationships, clearing fee relationships, and clearing timeliness relationships between bank nodes. The display module 200 is used to determine the target cross-border clearing path from the cross-border clearing path knowledge graph based on cross-border remittance information and to visualize the target cross-border clearing path. The number of target cross-border clearing paths is greater than or equal to 1. The head node of the target cross-border clearing path is the remitting bank of the cross-border remittance information, the tail node is the receiving bank of the cross-border remittance information, and the clearing currency of the banks corresponding to the intermediate nodes includes the clearing currency of the receiving bank. By constructing the cross-border clearing path knowledge graph, cross-border clearing paths are generated to meet the needs of cross-border clearing in various scenarios, thereby improving the level of cross-border financial services, solving the current difficulties faced by cross-border clearing, and improving the security and efficiency of cross-border transactions.

[0197] Another embodiment of this application provides a storage medium storing a computer program thereon, wherein the computer program, when executed by a processor, implements a cross-border clearing path recommendation method as described in any of the above embodiments.

[0198] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0199] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0200] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0201] Another embodiment of the present invention provides an electronic device, such as... Figure 11 As shown, it includes:

[0202] One or more processors 201.

[0203] Storage device 202, on which one or more programs are stored.

[0204] When one or more programs are executed by one or more processors 201, the one or more processors 201 implement the cross-border clearing path recommendation method as described in any of the above embodiments.

[0205] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts.

[0206] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

[0207] While several specific implementation details are included in the foregoing discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0208] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

[0209] The features described in the various embodiments of this specification can be substituted for or combined with each other. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The systems and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0210] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0211] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A cross-border clearing path recommendation method, characterized in that, The method comprises the following steps: constructing a cross-border clearing path knowledge graph according to data of multiple data sources, the multiple data sources at least comprising a SWIFT GPI, a CIPS, and a bank system, the cross-border clearing path knowledge graph at least comprising a clearing path relationship, a clearing cost relationship, and a clearing time limit relationship between bank nodes; determining a target cross-border clearing path from the cross-border clearing path knowledge graph according to cross-border remittance information, and visually displaying the target cross-border clearing path, the number of the target cross-border clearing path being greater than or equal to 1, a head node of the target cross-border clearing path being a remittance bank of the cross-border remittance information, a tail node being a receiving bank of the cross-border remittance information, and a clearing currency of an intermediate node corresponding to a bank comprising a clearing currency of the receiving bank. 2.The cross-border clearing path recommendation method of claim 1, wherein, After constructing the cross-border clearing path knowledge graph according to data of multiple data sources, the method further comprises the following steps: determining an update granularity of the cross-border clearing path knowledge graph; based on the update granularity, updating the cross-border clearing path knowledge graph by using data newly generated by the multiple data sources within the update granularity. 3.The cross-border clearing path recommendation method of claim 1, wherein, Constructing a cross-border clearing path knowledge graph according to data of multiple data sources comprises the following steps: cleaning data of the multiple data sources to obtain cross-border clearing data; extracting entities and relationships in the cross-border clearing data to obtain the cross-border clearing path knowledge graph, the entities comprising time entities, link entities, BIC entities, and composite BIC entities, and the relationships comprising transaction statistics relationships, solid line flow direction relationships, dotted line flow direction relationships, and currency flow direction relationships. 4.The cross-border clearing path recommendation method of claim 3, characterized in that, Cleaning data of the multiple data sources to obtain cross-border clearing data comprises the following steps: identifying error data in the data of the multiple data sources; filling in missing values in the error data and deleting abnormal values and repeated values in the error data. 5.The cross-border clearing path recommendation method of claim 1, wherein, Determining a target cross-border clearing path from the cross-border clearing path knowledge graph according to cross-border remittance information comprises the following steps: determining a query condition corresponding to the cross-border remittance information, the query condition at least comprising a remittance bank, a receiving bank, and a clearing currency of the cross-border remittance information; taking a cross-border clearing path in the cross-border clearing path knowledge graph satisfying the query condition as the target cross-border clearing path. 6.The cross-border payment route recommendation method of claim 1, wherein, Visually displaying the target cross-border clearing path comprises the following steps: determining a target visual display type of the target cross-border clearing path, the target visual display type being one of a commonly used clearing path recommendation display, a total clearing path recommendation display, a shortest clearing path recommendation display, a least cost clearing path recommendation display, a shortest time consumption clearing path display, and a risk level clearing path recommendation display; visually displaying the target cross-border clearing path according to the target visual display type. 7.A cross-border clearing path recommendation system, characterized in that, The method comprises the following steps: A construction module is configured to construct a cross-border clearing path knowledge graph according to data of a plurality of data sources, the plurality of data sources at least including SWIFT GPI, the Renminbi Cross-Border Payment System (CIPS), and a banking system, and the cross-border clearing path knowledge graph includes clearing path relationships, clearing cost relationships, and clearing time efficiency relationships between bank nodes. A display module is configured to determine a target cross-border clearing path from the cross-border clearing path knowledge graph according to cross-border remittance information, and to visually display the target cross-border clearing path, a number of the target cross-border clearing paths being greater than or equal to 1, a head node of the target cross-border clearing path being a remitting bank of the cross-border remittance information, a tail node being a receiving bank of the cross-border remittance information, and a target clearing currency of an intermediate node corresponding to a bank being a clearing currency of the receiving bank. 8.The cross-border clearing path recommendation system of claim 7, wherein, Further comprising: A determination module is configured to determine an update granularity of the cross-border clearing path knowledge graph. An update module is configured to update the cross-border clearing path knowledge graph based on the update granularity and using data newly generated by the plurality of data sources within the update granularity.

9. An electronic device, comprising: Comprise: One or more processors; A storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the cross-border clearing path recommendation method of any one of claims 1-6.

10. A storage medium, characterized by A computer program is stored thereon, and when the computer program is executed by a processor, the cross-border clearing path recommendation method of any one of claims 1-6 is implemented.

Images