Real-time financing process monitoring and early warning method and equipment based on knowledge graph

Abstract

The invention provides a financing process real-time monitoring and early warning method and device based on a knowledge graph, and the method comprises the steps: carrying out the real-time monitoring and early warning of a financing process according to multi-source target financial data which are collected in real time and / or at regular time and are generated in the financing process; dynamically updating a knowledge graph used for storing each node representing the type and attribute of the transaction entity and an association relationship between the nodes; abnormal index monitoring and transaction track data tracking are carried out on the knowledge graph in real time; determining an index financial risk early warning level corresponding to the abnormal index monitoring result data based on the customized financial risk rule data; and inputting the transaction trajectory data into a transaction anomaly detection model to obtain a trajectory financial risk score, and determining a trajectory financial risk level corresponding to the transaction trajectory data according to the trajectory financial risk score. According to the invention, the problems of difficulty in discovering hidden risks and poor adaptability caused by information fragmentation, insufficient monitoring and early warning real-time performance and weak association analysis capability of a traditional financing process monitoring mode can be solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of financial data processing, and in particular to a financing process real-time monitoring and early warning method and device based on a knowledge graph. BACKGROUND

[0002] In the financial field, a financing process involves multiple links (such as due diligence, risk assessment, contract signing, fund transfer, etc.), multiple participants (such as financing parties, financial institutions, regulatory authorities, etc.), and multiple types of data (such as financial data, credit data, operational data, etc.), and its compliance, security and efficiency directly affect the stability of the financial market. With the expansion of financing scale and the complication of the process, the traditional monitoring methods gradually expose their limitations.

[0003] Currently, financing process monitoring mainly relies on manual review, traditional database monitoring, single-link monitoring systems, and risk early warning based on fixed business rules, which leads to the problems of information fragmentation, insufficient real-time monitoring and early warning, weak correlation analysis capability, difficulty in discovering implicit risks, and poor adaptability in the existing financing process monitoring and early warning methods. SUMMARY

[0004] In view of this, the embodiments of the present application provide a financing process real-time monitoring and early warning method and device based on a knowledge graph to eliminate or improve one or more defects in the prior art.

[0005] One aspect of the present application provides a financing process real-time monitoring and early warning method based on a knowledge graph, comprising: dynamically updating a knowledge graph for storing nodes representing types and attributes of transaction entities and association relationships between the nodes according to multi-source target financial data generated in a financing process collected in real time and / or at regular intervals; performing real-time abnormal index monitoring on the knowledge graph to obtain corresponding abnormal index monitoring result data, and performing real-time transaction trajectory data tracking on the knowledge graph to obtain corresponding transaction trajectory data; determining an index financial risk early warning level corresponding to the abnormal index monitoring result data based on preset custom financial risk rule data, and inputting the transaction trajectory data into a preset transaction anomaly detection model to make the transaction anomaly detection model output a trajectory financial risk score corresponding to the transaction trajectory data, and determining a trajectory financial risk level corresponding to the transaction trajectory data according to the trajectory financial risk score.

[0006] In some embodiments of the present application, the financing process real-time monitoring and early warning method based on a knowledge graph further comprises: The data used to represent the financial risk warning level of the indicator and the financial risk level of the trajectory are pushed to the corresponding client devices respectively; The knowledge graph, the current financial risk warning level of the indicator, and the trajectory financial risk level are transmitted to a visualization interface for display; wherein, the visualization interface supports operations including zooming, filtering, and trajectory tracing of the knowledge graph; Based on the current financial risk warning level of the aforementioned indicators and the financial risk level of the aforementioned trajectory, corresponding warning report data that can be exported is generated.

[0007] In some embodiments of this application, before dynamically updating the knowledge graph used to store the types and attributes of each transaction entity and the relationships between the nodes based on multi-source target financial data generated in the financing process in real time and / or at regular intervals, the method further includes: Preprocessing is performed on the multi-source historical financial data corresponding to the financing process obtained from the internal and external systems of the target financial institution in advance, so as to extract the type, attributes and transaction relationship between each transaction entity and other transaction entities corresponding to the multi-source historical financial data; Each of the transaction entities is treated as a different node, and the transaction relationship between the different transaction entities connected by the nodes is represented by an edge, so as to construct a knowledge graph. Each node in the knowledge graph records the type and attribute of the transaction entity that is uniquely associated with it.

[0008] In some embodiments of this application, the types of the transaction entity include: financing entity, audit node, and fund account; The attributes corresponding to the financing entity are used to represent information such as the financing entity's name, registered capital, and credit rating. The attributes corresponding to the audit node are used to represent information such as the audit node name, the auditor, and the deadline. The attributes corresponding to the fund account are used to represent information such as the account number, the bank where the account was opened, and the balance.

[0009] In some embodiments of this application, the step of dynamically updating the knowledge graph used to store the types and attributes of each transaction entity and the relationships between the nodes based on multi-source target financial data generated during the financing process in real time and / or at regular intervals includes: If multi-source target financial data generated in real time during the financing process is collected from the internal system and / or external system corresponding to the target financial institution, the knowledge graph used to store the types and attributes of each transaction entity and the relationships between each node is updated in a triggered manner based on the multi-source target financial data. If the current time has reached the preset timed incremental update time, then extract the newly added non-real-time multi-source target financial data in the financing process, and incrementally update the knowledge graph based on the multi-source target financial data.

[0010] In some embodiments of this application, the real-time monitoring of anomaly indicators on the knowledge graph to obtain corresponding anomaly indicator monitoring result data; and the real-time tracking of transaction trajectory data on the knowledge graph to obtain corresponding transaction trajectory data, include: The multi-source target financial data is compared with the historical version of the knowledge graph before the update in real time to obtain the corresponding abnormal indicator monitoring results data; In addition, the knowledge graph is tracked in real time based on a preset graph traversal algorithm to obtain the corresponding transaction trajectory data, wherein the graph traversal algorithm includes: depth-first search.

[0011] In some embodiments of this application, the transaction anomaly detection model includes a graph neural network.

[0012] Another aspect of this application provides a knowledge graph-based real-time monitoring and early warning device for financing processes, including: The knowledge graph construction and update module is used to dynamically update the knowledge graph that stores the types and attributes of each transaction entity and the relationships between each node based on the multi-source target financial data generated in the financing process in real time and / or at regular intervals. The real-time monitoring module is used to monitor the knowledge graph for abnormal indicators in real time to obtain the corresponding abnormal indicator monitoring result data; and to track the knowledge graph for transaction trajectory data in real time to obtain the corresponding transaction trajectory data. The early warning analysis module is used to determine the indicator financial risk early warning level corresponding to the abnormal indicator monitoring result data based on preset custom financial risk rule data; and to input the transaction trajectory data into a preset transaction anomaly detection model so that the transaction anomaly detection model outputs the trajectory financial risk score corresponding to the transaction trajectory data, and determines the trajectory financial risk level corresponding to the transaction trajectory data based on the trajectory financial risk score.

[0013] A third aspect of this application provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the aforementioned knowledge graph-based real-time monitoring and early warning method for financing processes.

[0014] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned knowledge graph-based real-time monitoring and early warning method for financing processes.

[0015] The fifth aspect of this application provides a computer program product comprising a computer program that, when executed by a processor, implements the aforementioned knowledge graph-based real-time monitoring and early warning method for financing processes.

[0016] The knowledge graph-based real-time monitoring and early warning method for financing processes provided in this application dynamically updates a knowledge graph that stores nodes representing the types and attributes of various transaction entities and the relationships between these nodes, based on multi-source target financial data generated during the financing process in real time and / or at regular intervals. It monitors the knowledge graph for anomalies in real time to obtain corresponding anomaly monitoring results; tracks transaction trajectory data in the knowledge graph in real time to obtain corresponding transaction trajectory data; determines the financial risk warning level corresponding to the anomaly monitoring results based on preset custom financial risk rules; and inputs the transaction trajectory data into a preset transaction anomaly detection model, causing the model to output a trajectory financial risk score corresponding to the transaction trajectory data, and determines the trajectory financial risk level based on the score. This method effectively solves the technical problems of traditional financing process monitoring methods being unable to correlate multi-source data in real time and having difficulty identifying hidden risks in complex transaction networks through anomaly detection. Specifically, by constructing a unified knowledge graph, multi-source data in the financing process is integrated to break down information silos and achieve visualized association between entities and relationships; a real-time data access and graph update mechanism is developed to ensure the immediacy of monitoring data and dynamically track the status of the financing process; combined with the associative reasoning capabilities of the knowledge graph and intelligent algorithms, risk identification across stages and entities is achieved, improving the accuracy of early warnings; and dynamically adjustable early warning rules and self-learning models are designed to adapt to new risk patterns and reduce manual maintenance costs.

[0017] Additional advantages, objectives, and features of this application will be set forth in part in the description which follows, and will in part become apparent to those skilled in the art upon review of the following description, or may be learned by practice of the application. The objectives and other advantages of this application can be realized and obtained by means of the structures specifically pointed out in the specification and drawings.

[0018] Those skilled in the art will understand that the purposes and advantages that can be achieved with this application are not limited to those specifically described above, and that the above and other purposes that this application can achieve will be more clearly understood from the following detailed description. Attached Figure Description

[0019] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, do not constitute a limitation thereof. The components in the drawings are not drawn to scale but are merely for illustrating the principles of this application. For ease of illustration and description of certain parts of this application, corresponding portions in the drawings may be enlarged, i.e., may appear larger relative to other components in an exemplary device actually manufactured according to this application. In the drawings: Figure 1 This is a schematic diagram of the first process of a knowledge graph-based real-time monitoring and early warning method for financing processes in one embodiment of this application.

[0020] Figure 2 This is a schematic diagram of a second process of a knowledge graph-based real-time monitoring and early warning method for financing processes in one embodiment of this application.

[0021] Figure 3 This is a schematic diagram of the first structure of a knowledge graph-based real-time monitoring and early warning device for financing processes in one embodiment of this application.

[0022] Figure 4 This is a schematic diagram of a second structure of a knowledge graph-based real-time monitoring and early warning device for financing processes in one embodiment of this application.

[0023] Figure 5 This is a logical diagram illustrating the collaborative execution of a knowledge graph-based real-time monitoring and early warning method for financing processes by various modules in a knowledge graph-based real-time monitoring and early warning device, as described in an application example of this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the embodiments and accompanying drawings. Here, the illustrative embodiments and their descriptions are used to explain this application, but are not intended to limit it.

[0025] It should also be noted that, in order to avoid obscuring this application with unnecessary details, only the structures and / or processing steps closely related to the solution according to this application are shown in the accompanying drawings, while other details that are not closely related to this application are omitted.

[0026] It should be emphasized that the term "including / comprises" as used herein refers to the presence of a feature, element, step, or component, but does not exclude the presence or addition of one or more other features, elements, steps, or components.

[0027] It should also be noted that, unless otherwise specified, the term "connection" in this article can refer not only to a direct connection, but also to an indirect connection involving an intermediary.

[0028] In the following description, embodiments of the present application will be illustrated with reference to the accompanying drawings. In the drawings, the same reference numerals represent the same or similar parts, or the same or similar steps.

[0029] It should be noted that the monitoring of the financing process mainly relies on the following methods: Manual review: This involves manually checking paper or electronic documents and generating monitoring reports periodically. However, this method suffers from problems such as low efficiency, strong subjectivity, and poor real-time performance.

[0030] Traditional database monitoring stores financing data in relational databases and detects anomalies through SQL queries or simple rules (such as threshold judgments). However, it is difficult to handle complex relationships across stages and entities (such as implicit guarantees from related companies and nested fund flows).

[0031] Single-stage monitoring system: Specialized tools are developed for a specific stage of the financing process (such as pre-loan review), but the data of each system is isolated, forming "information silos" and making it impossible to achieve full-process linkage monitoring.

[0032] The rule engine provides early warnings based on fixed business rules (such as "early warning triggered when overdue for more than 30 days"), but the rules are updated late and it is difficult to deal with new risks (such as fraudulent trade financing, concealment of related-party transactions, etc.).

[0033] However, the above-mentioned financing process monitoring methods have the following problems: (1) Fragmented information: Financing data is scattered across multiple systems (such as CRM, ERP, and bank core systems), lacking unified connection, making it difficult to identify cross-stage risks (such as the same company raising funds repeatedly through different subsidiaries).

[0034] (2) Insufficient real-time performance: Relying on timed data synchronization or manual entry, it is impossible to capture dynamic changes in the process in real time (such as sudden abnormal transfer of funds), resulting in delayed early warning.

[0035] (3) Weak correlation analysis capability: Traditional databases and rule engines cannot model complex relationships between entities (such as corporate equity chains and guarantee networks), and it is difficult to discover hidden risks (such as multi-layered nested misappropriation of funds).

[0036] (4) Poor adaptability: Fixed rules are difficult to cover new risk patterns, and rule modification requires manual intervention, making it impossible to respond quickly to market changes.

[0037] Based on this, in order to solve the technical problems that traditional financing process monitoring methods cannot connect multi-source data in real time and that anomaly detection is difficult to identify hidden risks in complex transaction networks, this application provides a knowledge graph-based real-time monitoring and early warning method for financing processes, a knowledge graph-based real-time monitoring and early warning device for financing processes, physical equipment, computer-readable storage media, and computer program products for executing the knowledge graph-based real-time monitoring and early warning method. By dynamically updating the knowledge graph, multi-source data fusion is achieved, improving the real-time performance of monitoring; combined with a custom rule engine and machine learning model, it covers both explicit and implicit risk scenarios.

[0038] The following examples will provide a detailed description.

[0039] Based on this, embodiments of this application provide a knowledge graph-based real-time monitoring and early warning method for financing processes, which can be implemented by a knowledge graph-based real-time monitoring and early warning device for financing processes. See [link to relevant documentation]. Figure 1 The knowledge graph-based real-time monitoring and early warning method for financing processes specifically includes the following: Step 100: Based on the multi-source target financial data generated in the financing process collected in real time and / or at regular intervals, dynamically update the knowledge graph used to store the types and attributes of each transaction entity and the relationships between the nodes.

[0040] Understandably, the financing process refers to the entire business process for a company or individual, from initiating a financing need, submitting documents, reviewing and evaluating the application, signing a contract, to receiving funds and subsequent repayment.

[0041] In one or more embodiments of this application, the multi-source target financial data generated in the financing process refers to the financial data corresponding to the financing process monitored by the target financial institution, obtained from the internal system and / or external system corresponding to the target financial institution. This financial data can be real-time data collected in real time or non-real-time data collected periodically. The financial data can contain structured data and unstructured data. If it contains unstructured data, it needs to be preprocessed into structured data before dynamically updating the knowledge graph based on the unstructured data in step 100 to extract the type and attributes of the transaction entity and the relationship between the transaction entity and other transaction entities. If the financial data is structured data, the type and attributes of the transaction entity and the relationship between the transaction entity and other transaction entities can be extracted directly.

[0042] It should be noted that a knowledge graph is a semantic network based on a graph structure, composed of "entities" and "relationships," used to represent the relationships between different data, and can intuitively present the inherent connections between multi-dimensional information. In this embodiment, the knowledge graph is used to store each node and the relationships between them, wherein the nodes are used to record the type and attributes of transaction entities. The types of transaction entities may include financing entities, review nodes, and fund accounts, etc.

[0043] In this context, "relationship" can be abbreviated as "relationship" and refers to the connection between entities, such as "enterprise-submission-materials" or "reviewer-approval-contract". Nodes (or entities) are the basic elements that make up the knowledge graph, such as the financing entity (enterprise / individual), reviewer, fund account, and contract.

[0044] Step 200: Monitor the knowledge graph for abnormal indicators in real time to obtain the corresponding abnormal indicator monitoring results data; and track the knowledge graph for transaction trajectory data in real time to obtain the corresponding transaction trajectory data.

[0045] In step 200, the operational data and status changes of each stage of the financing process can be continuously and instantly tracked and recorded, and abnormal indicator monitoring results data and transaction trajectory data can be obtained respectively.

[0046] It should be noted that the abnormal indicator monitoring results data are used to indicate whether the knowledge graph obtained from the current monitoring contains abnormal indicators, and if it contains abnormal indicators, the specific type of the indicator, such as the type of abnormal indicator that a company's current financing amount exceeds the historical average by 300%.

[0047] The transaction trajectory data refers to key trajectories such as fund flow and review path, for example: "Financing entity A → Fund account A → Transfer to account B → Account B is a hidden account of related company C", etc.

[0048] Step 300: Based on the preset custom financial risk rule data, determine the indicator financial risk warning level corresponding to the abnormal indicator monitoring result data; and input the transaction trajectory data into the preset transaction anomaly detection model so that the transaction anomaly detection model outputs the trajectory financial risk score corresponding to the transaction trajectory data, and determine the trajectory financial risk level corresponding to the transaction trajectory data based on the trajectory financial risk score.

[0049] It should be noted that the early warning mechanism refers to a mechanism that automatically identifies and triggers warnings for abnormal data and risky behaviors that occur in the financing process based on preset rules or intelligent algorithms.

[0050] In step 300, the custom financial risk rule data can be user-defined, such as: (1) Rule 1: If "the related enterprises of the financing entity have a record of dishonesty" and "this financing is unsecured", a medium risk warning will be triggered; (2) Rule 2: If “the funds are transferred in full to a non-business account within 24 hours of receipt”, a high-risk warning will be triggered.

[0051] As described above, the knowledge graph-based real-time monitoring and early warning method for financing processes provided in this application integrates multi-source data in the financing process by constructing a unified knowledge graph, breaking down information silos, and realizing the visual association of entities and relationships; it develops a real-time data access and graph update mechanism to ensure the immediacy of monitoring data and dynamically track the status of the financing process; it combines the associative reasoning capabilities of the knowledge graph with intelligent algorithms to achieve cross-stage and cross-entity risk identification and improve the accuracy of early warnings; and it designs dynamically adjustable early warning rules and self-learning models to adapt to new risk patterns and reduce manual maintenance costs.

[0052] To further address the issue of risk warning results being scattered across different systems, making it difficult for decision-makers to obtain a timely overall view, this application provides a knowledge graph-based real-time monitoring and early warning method for financing processes. (See also...) Figure 2 The method for real-time monitoring and early warning of financing processes based on knowledge graphs further includes the following content after step 300: Step 400: Push the data representing the financial risk warning level of the indicator and the financial risk level of the trajectory to the corresponding client devices respectively.

[0053] Specifically, the system can automatically push information to the client devices of the responsible persons based on the risk level (low / medium / high), and high-risk warnings can simultaneously trigger related freezing operations (such as suspending fund transfers).

[0054] Step 500: Transmit the knowledge graph, the current indicator financial risk warning level, and the trajectory financial risk level to the visualization interface for display; wherein, the visualization interface supports operations including zooming, filtering, and trajectory tracing of the knowledge graph.

[0055] Specifically, it can provide a knowledge graph visualization interface to intuitively display the relationship between financing entities, fund flows, and review nodes, and supports operations such as zooming, filtering, and trajectory tracing.

[0056] Step 600: Generate corresponding early warning report data that can be exported based on the current financial risk warning level of the indicator and the trajectory financial risk level.

[0057] Specifically, it can generate real-time monitoring dashboards that display key indicators (such as the number of warnings per day and the distribution of high-risk processes) and support the export of warning reports. The real-time monitoring dashboard, through a dynamic and visualized knowledge graph of risk propagation paths and real-time warning indicators, serves as an intelligent decision-making hub for second-level location and collaborative handling of financing risks. Components can include a circular risk heatmap, dynamic indicator cards, and real-time alarm streams.

[0058] As can be seen from the above description, the knowledge graph-based real-time monitoring and early warning method for financing processes provided in this application can help to quickly locate risk sources by integrating early warning levels, knowledge graphs and visualization operations (such as trajectory tracing); and can effectively meet audit compliance requirements by automatically generating structured early warning reports.

[0059] To further address the problem that historical financial data is highly heterogeneous (e.g., unstructured contract texts) and fragmented, making it difficult to directly use for risk analysis, this application provides a knowledge graph-based real-time monitoring and early warning method for financing processes. (See also...) Figure 2 The method for real-time monitoring and early warning of financing processes based on knowledge graphs also includes the following content before step 100: Step 010: Preprocess the multi-source historical financial data corresponding to the financing process obtained from the internal and external systems of the target financial institution in advance, so as to extract the type, attributes and transaction relationships between the various transaction entities and other transaction entities corresponding to the multi-source historical financial data.

[0060] Specifically, it can perform OCR recognition and NLP semantic parsing on unstructured data (such as PDF reports) to extract entities (such as "company name" and "guarantee amount") and relationships (such as "XX company provides a guarantee for YY company").

[0061] The historical financial data related to the financing process obtained from the internal system of the target financial institution may include: CRM (Customer Relationship Management) system data (customer information), approval system data (approval records), and core system data (fund flow); the historical financial data related to the financing process obtained from the external system of the target financial institution may include: enterprise credit report (credit center interface), business registration information, and news and public opinion data, etc. (which can be obtained through authorized web crawling tools).

[0062] Step 020: Treat each of the transaction entities as different nodes, and use edges to represent the transaction relationships between the different transaction entities connected to them, so as to construct a knowledge graph. Each node in the knowledge graph records the type and attribute of the transaction entity that is uniquely associated with it.

[0063] To further address the lack of standardized modeling of entities across the entire financing process in traditional monitoring systems, this application provides a knowledge graph-based real-time monitoring and early warning method for financing processes. The types of transaction entities include: financing entities, review nodes, and fund accounts. The attributes corresponding to the financing entity are used to represent information such as the financing entity's name, registered capital, and credit rating. The attributes corresponding to the audit node are used to represent information such as the audit node name, the auditor, and the deadline. The attributes corresponding to the fund account are used to represent information such as the account number, the bank where the account was opened, and the balance.

[0064] To further address the issues of resource waste and delayed risk response caused by full knowledge graph updates, this application provides a knowledge graph-based real-time monitoring and early warning method for financing processes, see [link to relevant documentation]. Figure 2 Step 100 of the knowledge graph-based real-time monitoring and early warning method for financing processes specifically includes the following: Step 110: If multi-source target financial data generated in real time during the financing process is collected from the internal system and / or external system corresponding to the target financial institution, the knowledge graph used to store the types and attributes of each transaction entity and the relationships between the nodes is updated based on the multi-source target financial data.

[0065] Step 120: If the current time has reached the preset timed incremental update time, extract the newly added non-real-time multi-source target financial data in the financing process, and incrementally update the knowledge graph based on the multi-source target financial data.

[0066] Specifically, the triggered update process in step 100 may include: when the data source changes (e.g., the review status changes from "pending approval" to "approved"), updating the attributes or relationships of the corresponding entities in the graph in real time via the Webhook interface. The Webhook interface is a lightweight system integration mechanism based on HTTP callbacks, allowing applications to receive external event notifications in real time (rather than actively polling). The incremental update process in step 100 may include: incrementally loading non-real-time data (e.g., company annual reports) into the database every morning to avoid the resource consumption of a full update.

[0067] As can be seen from the above description, the knowledge graph-based real-time monitoring and early warning method for financing processes provided in this application ensures the real-time performance of key transactions (such as second-level response for large-amount transfers) through triggered updates; and reduces the processing overhead of non-real-time data (such as annual reports) through incremental updates.

[0068] To further address the issue that traditional data comparison requires a full table scan and cannot efficiently identify incremental anomalies, this application provides a knowledge graph-based real-time monitoring and early warning method for financing processes, see [link to relevant documentation]. Figure 2 Step 200 of the knowledge graph-based real-time monitoring and early warning method for financing processes specifically includes the following: Step 210: Compare the multi-source target financial data with the historical version of the knowledge graph before the update in real time to obtain the corresponding abnormal indicator monitoring results data.

[0069] Specifically, the current financing process data can be compared in real time with historical data in the knowledge graph and preset baselines (such as "average financing amount in the same industry") to monitor abnormal indicators (such as a company's financing amount suddenly increasing by 300% compared with the historical average).

[0070] And, step 220: Real-time tracking of transaction trajectory data of the knowledge graph based on a preset graph traversal algorithm to obtain the corresponding transaction trajectory data, wherein the graph traversal algorithm includes: depth-first search.

[0071] Specifically, key trajectories such as fund flow and review path can be tracked through graph traversal algorithms (such as depth-first search). For example: "Financing entity A → Fund account A → Transfer to account B → Account B is a hidden account of related company C".

[0072] To further address the challenge of traditional machine learning models in handling the topological features of graph-structured data, this application provides a real-time monitoring and early warning method for financing processes based on knowledge graphs. (See also...) Figure 2 The transaction anomaly detection model in the knowledge graph-based real-time monitoring and early warning method for financing processes can employ a graph neural network.

[0073] Anomaly detection models (such as the GNN neural network shown in the figure) can be trained in advance based on historical risk cases and graph features (such as entity relationship density and capital flow complexity).

[0074] Based on this, the core improvements of the knowledge graph-based real-time monitoring and early warning method for financing processes provided in this application are as follows: (1) Knowledge graph-based financing process entity and relationship modeling method: Define the entity types (such as "approval node" and "guarantee relationship") and attributes unique to the financing field to achieve semantic fusion of multi-source data.

[0075] (2) Real-time access to financing data and dynamic update mechanism of knowledge graph: Through trigger-based interface and incremental update algorithm, the graph data is synchronized with the financing process and supports second-level response.

[0076] (3) Combining knowledge graph multi-dimensional risk warning algorithm: integrating rule engine (known risks) and graph neural network (unknown risks) to achieve accurate identification of cross-linked risks.

[0077] (4) Financing Risk Visualization and Tracing System: The risk propagation path is displayed intuitively through a graph interface, supporting interactive analysis and decision support.

[0078] Based on this, in other embodiments of the knowledge graph-based real-time monitoring and early warning method for financing processes in this application, if the entity matching accuracy of the knowledge graph is insufficient, entity linking technology (such as semantic similarity calculation based on BERT) can be introduced to improve the accuracy of entity alignment across data sources; for non-core data (such as corporate annual reports), scheduled batch updates (such as once a day) can be used instead of triggered updates to reduce system resource consumption; if the deployment cost of machine learning models is high, it can be simplified to a hybrid mode of "knowledge graph and expert rule base", and basic early warning can be achieved by manually entering new risk rules; if the knowledge graph data volume is too large, a distributed graph database (such as a Neo4j cluster) can be used instead of single-machine storage to improve the processing capacity of large-scale data.

[0079] From a software perspective, this application also provides a knowledge graph-based real-time monitoring and early warning device for implementing all or part of the aforementioned knowledge graph-based real-time monitoring and early warning method for financing processes. See [link to relevant documentation]. Figure 3 The knowledge graph-based real-time monitoring and early warning device for financing processes specifically includes the following components: The knowledge graph construction and update module 10 is used to dynamically update the knowledge graph that stores the types and attributes of each transaction entity and the relationships between each node based on the multi-source target financial data generated in the financing process in real time and / or at regular intervals. The real-time monitoring module 20 is used to monitor the knowledge graph for abnormal indicators in real time to obtain the corresponding abnormal indicator monitoring result data; and to track the knowledge graph for transaction trajectory data in real time to obtain the corresponding transaction trajectory data. The early warning analysis module 30 is used to determine the indicator financial risk early warning level corresponding to the abnormal indicator monitoring result data based on preset custom financial risk rule data; and to input the transaction trajectory data into a preset transaction anomaly detection model so that the transaction anomaly detection model outputs the trajectory financial risk score corresponding to the transaction trajectory data, and determines the trajectory financial risk level corresponding to the transaction trajectory data based on the trajectory financial risk score.

[0080] The embodiments of the knowledge graph-based real-time monitoring and early warning device for financing processes provided in this application can be used to execute the processing flow of the embodiments of the knowledge graph-based real-time monitoring and early warning method for financing processes described above. Its functions will not be repeated here, but can be referred to the detailed description of the embodiments of the knowledge graph-based real-time monitoring and early warning method for financing processes described above.

[0081] The knowledge graph-based real-time monitoring and early warning device for financing processes can perform the real-time monitoring and early warning functions on either a server or a client device. The choice can be made based on the processing power of the client device and the limitations of the user's usage scenario. This application does not impose any limitations on this. If all operations are performed on the client device, the client device may further include a processor for the specific processing of the knowledge graph-based real-time monitoring and early warning of the financing process.

[0082] The aforementioned client device may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission with the server. The server may include a server on the task scheduling center side; in other implementation scenarios, it may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, a server cluster consisting of multiple servers, or a distributed server structure.

[0083] The server and the client device can communicate using any suitable network protocol, including those not yet developed as of the date of this application. Such network protocols may include, for example, TCP / IP, UDP / IP, HTTP, HTTPS, etc. Furthermore, such network protocols may also include RPC (Remote Procedure Call Protocol) and REST (Representational State Transfer Protocol) protocols used on top of the aforementioned protocols.

[0084] As described above, the knowledge graph-based real-time monitoring and early warning device for financing processes provided in this application integrates multi-source data in the financing process by constructing a unified knowledge graph, breaking down information silos and realizing the visual association of entities and relationships; it develops a real-time data access and graph update mechanism to ensure the immediacy of monitoring data and dynamically track the status of the financing process; it combines the associative reasoning capabilities of the knowledge graph with intelligent algorithms to achieve cross-stage and cross-entity risk identification and improve the accuracy of early warnings; and it designs dynamically adjustable early warning rules and self-learning models to adapt to new risk patterns and reduce manual maintenance costs.

[0085] In addition, the knowledge graph construction and updating module 10 in the knowledge graph-based real-time monitoring and early warning device for financing processes provided in this application embodiment is also used to execute step 020 in the above method embodiment; the early warning analysis module 30 can also be used to execute step 400. Based on this, a multi-source data acquisition module 01 can also be additionally set to execute step 010 in the above method embodiment; and, see also... Figure 4 Additionally, a visualization and interaction module 40 can be set up to execute steps 500 and 600 in the above method embodiments, which will be explained in detail in the following application examples.

[0086] To further illustrate the above embodiments, this application also provides a specific application example of a knowledge graph-based real-time monitoring and early warning method for financing processes executed by a knowledge graph-based real-time monitoring and early warning device, the purpose of which includes: (1) Construct a unified knowledge graph model, integrate multi-source data in the financing process, break down information silos, and realize the visual association between entities and relationships.

[0087] (2) Develop a real-time data access and map update mechanism to ensure the timeliness of monitoring data and dynamically track the status of the financing process.

[0088] (3) Combining the associative reasoning ability of knowledge graphs with intelligent algorithms, we can achieve risk identification across links and entities and improve the accuracy of early warning.

[0089] (4) Design dynamic early warning rules and self-learning models to adapt to new risk patterns and reduce manual maintenance costs.

[0090] Based on this, see Figure 5 The real-time monitoring and early warning method for financing processes based on knowledge graphs, executed by the aforementioned knowledge graph-based financing process real-time monitoring and early warning device, specifically includes the following: S1. Data Integration: Collect structured data (such as contract terms and financial statements) and unstructured data (such as due diligence reports and news and public opinion) from each stage of the financing process, and construct a knowledge graph that includes entities such as "financing entity", "approval node" and "funds account" and relationships such as "submission", "approval" and "association".

[0091] S2. Real-time updates: By accessing dynamic data from various systems in real time through interfaces (such as fund arrival notifications and audit status changes), the entity attributes and relationships in the graph are automatically updated to ensure data timeliness.

[0092] S3. Intelligent monitoring: Based on graph-based correlation reasoning (such as identifying the "actual controller" through equity relationships), combined with preset rules (such as "related enterprise guarantee amount exceeds 50% of net assets") and machine learning models (such as identifying abnormal fund flows), abnormal behavior is detected in real time.

[0093] S4. Tiered Early Warning: Different early warning methods (such as system prompts, SMS notifications, and manual intervention) are triggered according to the risk level (low / medium / high), and the risk transmission path is visualized through a graph (such as "Company A → Guarantee Company B → Funds flow to shell company C").

[0094] Specifically, the knowledge graph-based real-time monitoring and early warning device for financing processes includes the following components: 1. Multi-source data acquisition module, detailed description as follows: Function: Retrieves real-time financing-related data from internal and external data sources, including: (1) Internal data: financial institutions’ CRM system (customer information), approval system (audit records), core system (fund flow); (2) External data: Enterprise credit reports (credit center interface), business information, news and public opinion (web crawler tools).

[0095] (3) Data processing: Perform OCR recognition and NLP semantic parsing on unstructured data (such as PDF reports) to extract entities (such as "company name" and "guarantee amount") and relationships (such as "XX company provides guarantee for YY company").

[0096] 2. Knowledge graph construction and updating module, detailed explanation as follows: Graph modeling: (1) Entity layer: Define core entity types, such as “Financing entity” (attributes: name, registered capital, credit rating), “Review node” (attributes: node name, reviewer, deadline), and “Funds account” (attributes: account number, bank, balance).

[0097] (2) Relationship layer: Define the relationships between entities, such as "Financing entity - Submission - Application materials", "Review node - Approval - Financing application", "Financing entity - Related - Related enterprises".

[0098] Real-time update mechanism: (1) Triggered update: When the data source changes (such as the review status changing from "pending approval" to "approved"), the attributes or relationships of the corresponding entities in the graph are updated in real time through the Webhook interface.

[0099] (2) Incremental update: Incremental data (such as corporate annual reports) are added to the database every day at midnight to avoid the resource consumption of full update.

[0100] 3. Real-time monitoring module, detailed description as follows: (1) Real-time comparison: The current financing process data is compared with the historical data in the knowledge graph and the preset baseline (such as "average financing amount in the same industry") in real time to monitor abnormal indicators (such as a company's financing amount suddenly increasing by 300% compared with the historical average).

[0101] (2) Related trajectory tracking: Track key trajectories such as fund flow and audit path through graph traversal algorithms (such as depth-first search). For example: "Financing entity A → Fund account A → Transfer to account B → Account B is a hidden account of related enterprise C".

[0102] 4. Early warning analysis module, detailed description as follows: Rules engine: Built-in configurable rule library, supporting user-defined risk rules, such as: (1) Rule 1: If "the related enterprises of the financing entity have a record of dishonesty" and "this financing is unsecured", a medium risk warning will be triggered; (2) Rule 2: If “the funds are transferred in full to a non-business account within 24 hours of receipt”, a high-risk warning will be triggered.

[0103] Machine learning models: (1) Training phase: Based on historical risk cases (such as financing fraud) and graph features (such as entity relationship density and capital flow complexity), train anomaly detection models (such as the GNN neural network shown in the figure).

[0104] (2) Reasoning stage: Input the current map data in real time, and the model outputs the risk probability. When the probability exceeds the threshold, an early warning is triggered.

[0105] (3) Early warning classification: Based on the risk level (low / medium / high), the warning will be automatically pushed to the corresponding responsible person. High risk warning will trigger the freezing of related operations (such as suspending fund transfer).

[0106] 5. Visualization and Interaction Module, detailed explanation as follows: Generate a real-time monitoring dashboard that displays key indicators (such as the number of warnings per day and the distribution of high-risk processes), and supports exporting warning reports.

[0107] In other words, the application example in this application uses knowledge graph technology to achieve real-time monitoring and early warning of the financing process, and specifically includes the following beneficial effects: (1) More comprehensive information integration: By linking multi-source data through knowledge graphs, the “information silos” can be broken down, and hidden risks across links and entities (such as chain guarantees by related enterprises) can be identified, solving the “fragmentation” problem of traditional databases.

[0108] (2) Significantly improved real-time performance: Based on the trigger-based update mechanism, the data delay is controlled at the second level, which can capture dynamics such as abnormal fund transfers and changes in audit status in real time, overcoming the lag of manual monitoring and timed synchronization.

[0109] (3) Stronger association analysis capability: By utilizing the reasoning ability of knowledge graphs, multi-level relationships (such as "actual controller of the enterprise → related enterprises → final flow of funds") can be traced, and nested risks that traditional rule engines cannot identify can be discovered.

[0110] (4) Higher early warning adaptability: Combining rule engine and machine learning model, it can not only cover known risks, but also identify new risks (such as new trade finance fraud) through self-learning, reducing the cost of manual rule maintenance.

[0111] (5) More intuitive visualization: The map visualization interface makes the risk transmission path clear at a glance, which makes it easier for decision-makers to quickly locate the problem links and improve the efficiency of emergency response.

[0112] This application also provides an electronic device, which may include a processor, a memory, a receiver, and a transmitter. The processor is used to execute the knowledge graph-based real-time monitoring and early warning method for financing processes mentioned in the above embodiments. The processor and memory can be connected via a bus or other means, taking a bus connection as an example. The receiver can be connected to the processor and memory via wired or wireless means.

[0113] The processor can be a central processing unit (CPU). The processor can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations of the above types of chips.

[0114] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the knowledge graph-based real-time monitoring and early warning method for financing processes in the embodiments of this application. The processor executes various functional applications and data processing by running the non-transitory software programs, instructions, and modules stored in the memory, thereby realizing the knowledge graph-based real-time monitoring and early warning method for financing processes in the above method embodiments.

[0115] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created by the processor, etc. Furthermore, the memory may include high-speed random access memory and non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0116] The one or more modules are stored in the memory, and when executed by the processor, the knowledge graph-based financing process real-time monitoring and early warning method in the embodiment is executed.

[0117] In some embodiments of this application, the user equipment may include a processor, a memory, and a transceiver unit. The transceiver unit may include a receiver and a transmitter. The processor, memory, receiver, and transmitter may be connected via a bus system. The memory is used to store computer instructions, and the processor is used to execute the computer instructions stored in the memory to control the transceiver unit to send and receive signals.

[0118] As one implementation method, the functions of the receiver and transmitter in this application can be implemented by transceiver circuits or dedicated transceiver chips, and the processor can be implemented by dedicated processing chips, processing circuits or general-purpose chips.

[0119] As another implementation approach, the server provided in this application embodiment can be implemented using a general-purpose computer. That is, the program code implementing the processor, receiver, and transmitter functions is stored in memory, and the general-purpose processor implements the processor, receiver, and transmitter functions by executing the code in memory.

[0120] This application also provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the steps of the aforementioned knowledge graph-based real-time monitoring and early warning method for financing processes. The computer-readable storage medium can be a tangible storage medium, such as random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, floppy disks, hard disks, removable storage disks, CD-ROMs, or any other form of storage medium known in the art.

[0121] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the aforementioned knowledge graph-based real-time monitoring and early warning method for financing processes.

[0122] Those skilled in the art will understand that the exemplary components, systems, and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Whether 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 implementation should not be considered beyond the scope of this application. When implemented in hardware, it can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. The programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave.

[0123] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0124] In this application, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or in place of features of other embodiments.

[0125] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to the embodiments of this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A knowledge graph-based financing process real-time monitoring and early warning method, characterized in that, The method comprises the following steps: According to the multi-source target financial data generated in the financing process collected in real time and / or at regular intervals, dynamically update the knowledge graph for storing nodes representing the types and attributes of each transaction entity and the association relationship between each node; Real-time monitoring of the knowledge graph for abnormal indicators to obtain corresponding abnormal indicator monitoring result data; And, real-time tracking of the knowledge graph for transaction trajectory data to obtain corresponding transaction trajectory data; Based on the pre-set custom financial risk rule data, determine the index financial risk warning level corresponding to the abnormal indicator monitoring result data; And input the transaction trajectory data into the pre-set transaction anomaly detection model to make the transaction anomaly detection model output the trajectory financial risk score corresponding to the transaction trajectory data, and determine the trajectory financial risk level corresponding to the transaction trajectory data according to the trajectory financial risk score. 2.The knowledge graph-based financing process real-time monitoring and early warning method according to claim 1, characterized in that, It also includes: Push the data representing the index financial risk warning level and the trajectory financial risk level to the corresponding client device respectively; The knowledge graph, the current index financial risk warning level and the trajectory financial risk level are transmitted to the visualization interface for display; wherein the operations supported by the visualization interface include: zooming, filtering and trajectory tracking of the knowledge graph; According to the current index financial risk warning level and the trajectory financial risk level, generate corresponding warning report data that supports export. 3.The knowledge graph-based financing process real-time monitoring and early warning method according to claim 1, characterized in that, Before the multi-source target financial data generated in the financing process collected in real time and / or at regular intervals is used to dynamically update the knowledge graph for storing nodes representing the types and attributes of each transaction entity and the association relationship between each node, it also includes: Pre-process the multi-source historical financial data corresponding to the financing process obtained from the internal and external systems of the target financial institution in advance, to extract the types, attributes and transaction relationships with other transaction entities of each transaction entity corresponding to the multi-source historical financial data; Each of the transaction entities is used as a different node, and the transaction relationship between the different transaction entities connected by the edge is used to construct a knowledge graph, and each node in the knowledge graph records the type and attribute of the corresponding transaction entity. 4.The knowledge graph-based financing process real-time monitoring and early warning method according to claim 3, characterized in that, The types of the transaction entities include: financing subjects, audit nodes and fund accounts; The attributes corresponding to the financing subject are used to represent the information of the financing subject name, registered capital and credit rating; The attributes corresponding to the audit node are used to represent the information of the audit node name, auditor and deadline; The attributes corresponding to the fund account are used to represent the information of the account number, opening bank and balance. 5.The knowledge graph-based financing process real-time monitoring and early warning method according to claim 1, characterized in that, The method comprises the following steps: If the current internal system and / or external system corresponding to the target financial institution collects multi-source target financial data generated in real time in the financing process, the knowledge graph for storing nodes representing types and attributes of transaction entities and association relationships between the nodes is triggered to update based on the multi-source target financial data; If the current time has reached the preset timing incremental update time, the newly added non-real-time multi-source target financial data of the financing process is extracted, and the knowledge graph is incrementally updated based on the multi-source target financial data. 6.The knowledge graph-based financing process real-time monitoring and early warning method according to claim 1, characterized in that, The knowledge graph is monitored in real time to obtain corresponding abnormal index monitoring result data; And, the knowledge graph is tracked in real time to obtain corresponding transaction trajectory data, including: The multi-source target financial data is compared with a historical version of the knowledge graph before the update in real time to obtain corresponding abnormal index monitoring result data; And, the knowledge graph is tracked in real time based on a preset graph traversal algorithm to obtain corresponding transaction trajectory data, wherein the graph traversal algorithm includes depth-first search. 7.The knowledge graph-based financing process real-time monitoring and early warning method according to claim 1, characterized in that, The transaction anomaly detection model includes a graph neural network.

8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the financing process real-time monitoring and early warning method based on the knowledge graph according to any one of claims 1 to 7.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the financing process real-time monitoring and early warning method based on the knowledge graph according to any one of claims 1 to 7.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to realize the financing process real-time monitoring and early warning method based on the knowledge graph according to any one of claims 1 to 7.

Images