Financial risk analysis method and device, equipment, storage medium and program product

Abstract

The embodiment of the invention provides a financial risk analysis method and device, equipment, a storage medium and a program product, and relates to the field of financial science and technology. The method comprises the steps of performing data fusion processing on unstructured data and structured data to obtain fused data; updating the knowledge graph based on the fused data; and performing risk analysis on the updated knowledge graph through a graph calculation algorithm to obtain a visual risk analysis result, the risk analysis result comprising a risk conduction path. According to the method, the analysis and evaluation capability of financial risks is enhanced.

Description

Technical Field

[0001] This application relates to the field of financial technology, and in particular to a financial risk analysis method, apparatus, device, storage medium, and program product. Background Technology

[0002] In modern banking, financial risk control is a core element in ensuring the safety of credit assets. With economic globalization and the increasing complexity of financial operations, banks face increasingly diverse types of risks.

[0003] Traditional risk control methods mainly rely on static structured data (such as corporate financial statements and historical credit records) and human experience judgment, which have a weak ability to analyze and assess financial risks. Summary of the Invention

[0004] This application provides a financial risk analysis method, apparatus, equipment, storage medium, and product to address the technical problem of weak financial risk analysis and assessment capabilities.

[0005] In a first aspect, this application provides a financial risk analysis method, including: performing data fusion processing on unstructured data and structured data to obtain fused data; updating a knowledge graph based on the fused data; and performing risk analysis on the updated knowledge graph using a graph computing algorithm to obtain visualized risk analysis results, the risk analysis results including risk transmission paths.

[0006] In one possible implementation of the first aspect, updating the knowledge graph based on fused data includes: performing entity recognition and entity relation extraction on the fused data to obtain a first triplet; creating a second triplet in the knowledge graph based on the first triplet, or updating the attribute information of a third triplet in the knowledge graph based on the first triplet, wherein the entities and entity relations of the third triplet are the same as those of the first triplet.

[0007] In one possible implementation of the first aspect, entity recognition and entity relationship extraction are performed on the fused data to obtain the first triplet, including: performing entity recognition and entity relationship recognition on the fused data through a deep learning model to obtain entity recognition results; and correcting the entity recognition results based on a logical rule library through a rule engine to obtain the first triplet.

[0008] In one possible implementation of the first aspect, data fusion processing is performed on unstructured data and structured data to obtain fused data, including: acquiring unstructured data and structured data; and performing data fusion processing on unstructured data and structured data to obtain fused data.

[0009] In one possible implementation of the first aspect, risk transmission path analysis is performed on the updated knowledge graph using a graph computation algorithm to obtain risk analysis results. This includes: determining the corresponding first graph computation algorithm and first graph computation parameters based on the risk scenario type; determining the corresponding first graph computation strategy based on the specifications of the knowledge graph; and performing risk analysis on the knowledge graph using the first graph computation algorithm based on the first graph computation parameters and the first graph computation strategy to obtain risk analysis results.

[0010] In one possible implementation of the first aspect, the data fusion processing of unstructured data and structured data to obtain fused data includes: performing data cleaning and data format standardization processing on unstructured data to obtain processed unstructured data; and performing data fusion processing on the processed unstructured data and structured data to obtain fused data.

[0011] In one possible implementation of the first aspect, the risk analysis results also include risk data sources. The risk transmission path is analyzed on the updated knowledge graph using a graph computing algorithm to obtain the risk analysis results, including: determining the risk data sources corresponding to the risk transmission path.

[0012] In one possible implementation of the first aspect, the graph database corresponding to the knowledge graph is a distributed graph database.

[0013] In one possible implementation of the first aspect, the method further includes: receiving a first query request; determining a first graph data sub-database corresponding to the first query request through a sharding strategy, wherein the first graph data sub-database includes the structured data corresponding to the first query request, and the multiple graph data sub-databases corresponding to the graph database include the first graph data sub-database.

[0014] Secondly, this application provides a financial risk analysis device, comprising:

[0015] The data fusion module is used to perform data fusion processing on unstructured and structured data to obtain fused data;

[0016] The update module is used to update the knowledge graph based on the fused data;

[0017] The risk analysis module is used to perform risk analysis on the updated knowledge graph using graph computing algorithms, and obtain visualized risk analysis results, including risk transmission paths.

[0018] Thirdly, this application provides an electronic device, including: a processor and a memory communicatively connected to the processor;

[0019] The memory stores the instructions that the computer executes;

[0020] The processor executes computer-executable instructions stored in memory to implement any of the methods of the first aspect.

[0021] Fourthly, this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method of any one of the first aspects.

[0022] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the method of any one of the first aspects.

[0023] The financial risk analysis method, apparatus, equipment, storage medium, and products provided in this application can update knowledge graphs by fusing data and perform risk analysis on the updated knowledge graphs using graph computing algorithms to obtain visualized risk analysis results. By performing risk analysis based on external unstructured data and internal structured data, the ability to analyze and assess financial risks is enhanced. On the other hand, by deeply mining complex relationships through graph computing algorithms, risk transmission paths are obtained, which can predict potential risk transmission problems. Furthermore, the visualized risk transmission paths can improve the interpretability of risk analysis results. Attached Figure Description

[0024] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0025] Figure 1 A schematic diagram illustrating a scenario for a financial risk analysis method provided in an embodiment of this application;

[0026] Figure 2 A flowchart illustrating a financial risk analysis method provided in this application embodiment;

[0027] Figure 3 A flowchart illustrating another financial risk analysis method provided in this application embodiment;

[0028] Figure 4 A schematic diagram of the structure of a financial risk analysis device provided in this application embodiment;

[0029] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0030] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0031] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0032] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of the relevant data all comply with the relevant laws, regulations, and standards of the relevant countries and regions, have taken necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation access points for users to choose to authorize or refuse.

[0033] Furthermore, the technical solution involved in this application, which involves big data analysis of user information (including but not limited to personal biometrics, identity data, consumption data, asset data, electronic terminal operation data, etc.) and the use of artificial intelligence technology for automated decision-making, and makes decisions that have a significant impact on personal rights based on the results of automated decision-making, provides users with corresponding operation entry points for users to choose to agree to or reject the results of automated decision-making; if the user chooses to reject, the process will proceed to the expert decision-making process.

[0034] It should be noted that the financial risk analysis methods, devices, equipment, storage media and program products provided in this application can be used in the field of financial technology, or in any field other than financial technology. The application fields of the financial risk analysis methods, devices, equipment, storage media and program products in this application are not limited.

[0035] The specific application scenario for this application can be the risk analysis and assessment of customers by banks. Figure 1 This is a schematic diagram illustrating an application scenario of a financial risk analysis method provided in an embodiment of this application. For example... Figure 1 As shown, bank risk control personnel can input relevant operation instructions through electronic devices to perform risk analysis and assessment of customers. After receiving the relevant instructions from the bank's risk control personnel, the electronic device can perform risk analysis and assessment of the customer.

[0036] The electronic device can communicate with a server cluster, which includes at least one server storing structured data. It can also communicate with external networks, enabling it to perform risk analysis and assessment based on both structured and unstructured data from the network.

[0037] In some embodiments, the electronic device may further include a display module, which can display the risk analysis results obtained from the risk analysis assessment to the bank's risk control personnel.

[0038] The electronic device in the embodiments of this application may be a computer, smartphone, tablet computer or other form of electronic device, and is not specifically limited here.

[0039] The specific application scenarios of the financial risk analysis method provided in this application can be as follows: When banks approve corporate loans, they need to comprehensively assess the company's financial condition, historical credit record, guarantee relationships, industry risks, etc., to identify potential risks and formulate reasonable credit granting strategies. Alternatively, after loan disbursement, banks need to continuously monitor the dynamic risks of borrowing companies, such as changes in the operations of upstream and downstream companies in the supply chain, and the impact of policy adjustments on the industry, to provide timely warnings of potential default risks. It can also be used when customers are conducting transactions, by analyzing customer transaction flows, related party relationships, and fund flow paths to identify abnormal transaction risks. Furthermore, it can be used to dynamically track risk transmission paths for complex relationships such as guarantee chains, equity chains, and capital chains in financial networks, preventing the spread of risks between financial institutions or enterprises. In addition, it can be applied to other specific scenarios, which are not limited here.

[0040] In modern banking, financial risk control is a core element in ensuring the safety of credit assets. With economic globalization and the increasing complexity of financial operations, banks face increasingly diverse types of risks.

[0041] Traditional risk control methods mainly rely on static structured data (such as corporate financial statements and historical credit records) and human experience judgment, which have a weak ability to analyze and assess financial risks.

[0042] For example, existing risk control technologies mainly rely on one or more of the following solutions:

[0043] 1. A scoring model based on structured data.

[0044] Credit scores are assigned to customers using statistical models based on the bank's internal historical and structured data, such as the company's financial statements and historical credit records.

[0045] However, the credit score in this scheme relies heavily on financial statements and historical data as "lagging indicators," making it slow to react to sudden, real-time external events (such as the bankruptcy of key suppliers or sudden changes in industry policies). It can only reflect the customer's historical performance and is slow to respond to sudden risk events.

[0046] 2. Manual monitoring of public opinion and due diligence.

[0047] Risk control personnel manually collect information about target companies through search engines or public channels as a supplementary reference for scoring.

[0048] However, this approach heavily relies on human experience to assess risk, resulting in low efficiency and limited coverage. It struggles to penetrate complex networks of relationships, such as equity, guarantees, and supply chains. It also fails to uncover hidden related-party risks and risk transmission paths. For example, it cannot quickly identify that Company B, an upstream supplier of Company A applying for a loan, happens to be a major guarantor for another bad debt transaction, Company C.

[0049] 3. Internal business system analysis.

[0050] The system analyzes the client's financial risks by using different risk management systems within the system and systems corresponding to different business processes.

[0051] However, in this scheme, each system conducts financial risk analysis independently, and the relevant data and analysis results are not shared between systems, resulting in low data utilization and the formation of information silos.

[0052] In summary, existing risk control solutions have relatively weak capabilities in analyzing and assessing financial risks.

[0053] The financial risk analysis methods, apparatus, equipment, storage media, and program products provided in this application are intended to solve the aforementioned technical problems of the prior art.

[0054] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0055] Figure 2 A flowchart illustrating a financial risk analysis method provided in this application embodiment is shown below. Figure 2 As shown, the method includes:

[0056] In this embodiment of the application, the electronic device performs financial risk analysis in response to a user's operation instruction, such as a first query request. Therefore, optionally, the method 200 may include the following steps:

[0057] S201. The electronic device receives the first query request and obtains unstructured data and structured data.

[0058] Upon receiving the first query request, the electronic device can obtain the unstructured data and structured data corresponding to the first query request.

[0059] In this embodiment of the application, the first query request is used to instruct the analysis of the relevant risks of the first entity, so the electronic device can first obtain the unstructured data and structured data related to the first entity.

[0060] In this embodiment of the application, the unstructured data refers to a collection of (unstructured) data from sources outside the bank and in various formats, such as news reports and legal documents related to a certain entity.

[0061] In this embodiment of the application, the structured data includes structured data from within the bank (such as credit systems and financial systems), such as customer information and financial statements.

[0062] In one possible implementation, the graph database corresponding to the knowledge graph is a distributed graph database. After receiving the first query request, the electronic device can determine the first graph data sub-database corresponding to the first query request through a sharding strategy. The first graph data sub-database includes the structured data corresponding to the first query request, and the multiple graph data sub-databases corresponding to the graph database include the first graph data sub-database.

[0063] In this embodiment of the application, the sharding strategy can be a hash based on node identifier, a hash based on node ID, or a graph partitioning algorithm. In addition, it can be other algorithms, which are not limited here.

[0064] In other words, the database storing structured data is a distributed database. When an electronic device processes a first query request, it can first determine the corresponding graph data sub-database for the first query request. Since the first graph data sub-database stores the structured data corresponding to the first query request, the first query request can be assigned to the first graph data sub-database for processing.

[0065] In this possible implementation method, in storage and query optimization, the system uses a consistent hashing algorithm to load balance query requests. For example, requests can be routed to the shards of storage enterprise A and its associated nodes first, reducing cross-node communication overhead. This allows complex graph analysis (such as multi-hop path discovery) to maintain a high response speed in large-scale graphs, thereby supporting the real-time risk warning requirements.

[0066] In one possible implementation, an electronic device can obtain structured data corresponding to a certain entity by accessing internal bank structured data through a database interface. This database interface can be connected to the bank's internal core business system or credit system, which includes the structured data related to the entity.

[0067] In one possible implementation, an electronic device can obtain unstructured data corresponding to an entity by using an application programming interface (API) connector. For example, the API connector can interface with third-party data providers such as business information and judicial information to obtain business information or judicial information related to the entity.

[0068] In one possible implementation, electronic devices acquire unstructured data corresponding to a specific entity, and can also use a web crawler cluster to collect publicly available unstructured data from the internet (news portals or industry forums). For example, news related to the entity can be obtained from a news portal.

[0069] For example, after risk control personnel input "Company A" as the analysis target into the system interface of an electronic device, the device can, on the one hand, obtain structured data such as financial information and credit history of Company A from the bank's internal database through a database interface. On the other hand, it can launch a web crawler to retrieve recent publicly available news information about Company A and its related parties (such as Company A's legal representative, shareholders, and core suppliers). It can also connect with third-party data providers such as business registration information and judicial information through API connectors to obtain relevant business registration information or judicial information about the entity. This information from external data sources, such as publicly available news information, business registration information, and judicial information, is all unstructured data.

[0070] S202. Electronic equipment performs data fusion processing on unstructured data and structured data to obtain fused data.

[0071] In one possible implementation, the electronic device can preprocess both unstructured and structured data before performing data fusion processing.

[0072] For example, electronic devices can perform data cleaning and data format standardization on unstructured data to obtain processed unstructured data; then, data fusion processing is performed on the processed unstructured data and structured data to obtain fused data.

[0073] This data cleaning can involve cleaning up data noise in unstructured data that is irrelevant to risk analysis. For example, it can remove hypertext markup language (HTML) tags, duplicate content, and other content from unstructured data.

[0074] The data format standardization process can be used to process different data formats of some unstructured data into the same format to facilitate subsequent processing. For example, timestamp data in different formats (e.g., "2023-05-01" and "05 / 01 / 2023") can be unified into the same format.

[0075] This data fusion process can be a process of cleaning, standardizing, and structuring multi-source data to ensure data consistency and availability.

[0076] In one possible implementation, the electronic device can be a centralized device cluster, which may include edge service nodes. When preprocessing unstructured and structured data, the electronic device can first preprocess the unstructured and structured data at the edge node devices, and then the edge node transmits the preprocessed data to the central node.

[0077] In some embodiments, the graph database corresponding to the knowledge graph is a distributed graph database, with a corresponding graph data sub-database deployed on the edge node device. After the edge node device preprocesses the unstructured data, it can store the preprocessed data in the graph data sub-database.

[0078] In this possible implementation, before data fusion processing, heterogeneous data can be preliminarily cleaned and formatted through edge computing nodes, thereby reducing the computing pressure on the central server, shortening the data fusion time interval, improving the efficiency of data fusion, and increasing the update speed of the knowledge graph.

[0079] In one possible implementation, the electronic device can fuse structured and unstructured data by associating unstructured data with structured data. For example, for a certain entity, the structured data inside the bank includes the entity's first attribute information, and the unstructured data outside the bank also includes the entity's second attribute information. The first device can fuse these two parts of attribute information, remove irrelevant information in the unstructured data, and also remove information that is redundant in the unstructured and structured data regarding the entity's attributes, thereby obtaining the entity's comprehensive and fused attribute information.

[0080] It is understood that the fusion of structured and unstructured data in the embodiments of this application is the integration of relevant information on the same target (the attributes of an entity and / or the relationship between an entity and other entities). Therefore, it can also be achieved in other ways, which are not limited here.

[0081] S203. Electronic devices update knowledge graphs based on fused data.

[0082] After merging structured and unstructured data to obtain fused data, electronic devices can update the knowledge graph based on this fused data.

[0083] In this embodiment, the knowledge graph is a large-scale semantic network whose basic unit is the "entity-relationship-entity" triple. It is not a traditional database, but a knowledge base used to represent entities in the real world and the complex relationships between them.

[0084] In one possible implementation, such as Figure 3 As shown, the electronic device can first identify the triples corresponding to the fused data. Then, the electronic device can determine whether there is a triple in the knowledge graph that corresponds to the fused data.

[0085] In this embodiment of the application, determining whether there is a triple in the knowledge graph that corresponds to the fused data can be done by determining whether there is a triple in the knowledge graph that is identical to the triple generated based on the fused data in terms of both entity and entity relationship.

[0086] If a corresponding triple exists, the electronic device does not need to create a new triple in the knowledge graph; it only needs to update the relevant information of the matching triple in the knowledge graph. If a corresponding triple does not exist in the knowledge graph, the electronic device can create a matching triple in the knowledge graph based on the first triple.

[0087] For example, the electronic device can first perform entity recognition and entity relationship extraction on the fused data to obtain the first triplet.

[0088] If no corresponding triple exists in the knowledge graph for the first triple, the electronic device can create a second triple in the knowledge graph based on the first triple. This second triple is identical to the first triple. That is, the two entities, the relationship between the two entities, and the attribute information of the two entities are all the same in the first and second triples.

[0089] If a triple corresponding to the first triple exists in the knowledge graph, the electronic device can update the attribute information of the third triple in the knowledge graph based on the first triple. The entities and entity relationships of the third triple are the same as those of the first triple. That is, the subject entity, object entity, and the relationship between the subject and object entities in the first triple are the same. However, at least one of the following three attributes—subject entity, object entity, and the relationship between the subject and object entities—is different between the first and third triples. Therefore, the electronic device can update the attribute information of the third triple in the knowledge graph based on the first triple, such as adding news sources or updating timestamps.

[0090] In this embodiment, when updating the attribute information of the third triplet in the knowledge graph based on the first triplet, the corresponding attribute information of the two can be merged or overwritten. For conflicting attribute information, the information with the latest timestamp can be determined as the final information. Since the first triplet has stronger real-time performance, in the event of attribute information conflict, the attribute information of the first triplet can be used by default.

[0091] For example, an electronic device identifies the entities "Company A", "Company C", and "major environmental litigation" from the news article "Company A's supplier Company C is involved in a major environmental lawsuit," and extracts the "supplier" and "involved" relationships. The system then compares the generated triples (Company A, Supplier, Company C) and (Company C, involved, major environmental litigation) with existing triples in the knowledge graph. For instance, it can compare entity node IDs, i.e., query whether there is an entity node corresponding to Company C in the knowledge graph. If a corresponding triple exists, the relevant attribute information of the entity "Company C" in the triple is updated, adding the information that Company C is involved in a major environmental lawsuit. If no corresponding triple exists, the electronic device creates new triples (Company A, Supplier, Company C) and (Company C, involved, major environmental litigation) in the knowledge graph.

[0092] It is understood that, in the embodiments of this application, creating a new triple in the knowledge graph is essentially creating a new node and / or a new edge in the graph. For example, when an electronic device creates a new triple (Company A, Supplier, Company C) or (Company C, Involved, Major Environmental Litigation) in the knowledge graph, it can create nodes and attributes (such as industry), and then create a "Supplier" relationship edge from "Company A" to "Company C", and a "Involved" relationship edge from "Company C" to "Major Environmental Litigation".

[0093] In this possible implementation, entity recognition and relation extraction technologies enable the efficient transformation of unstructured data into structured knowledge. On the other hand, by integrating data to update the knowledge graph, the dynamic updating capability of the knowledge graph is ensured. The dynamic updating mechanism ensures that the knowledge graph always reflects the latest risk dynamics, improves the real-time nature of the data used in risk analysis, and thus improves the accuracy of financial risk analysis results.

[0094] In one possible implementation, after the electronic device performs entity recognition and entity relationship extraction on the fused data to obtain the first triplet, it can use a deep learning model to perform entity recognition and entity relationship recognition on the fused data to obtain the entity recognition result, and then generate the corresponding triplet (i.e. the first triplet) based on the entity recognition result.

[0095] In this embodiment, the deep learning model can be a neural network-based model that can be used to automatically identify entities and relationships in text. For example, the deep learning model can be a natural language processing (NLP) engine based on deep learning, which may include a named entity recognition (NER) model, a relation extraction (RE) model, and an event extraction (EE) model. Taking a news article as an example, the electronic device can first identify entities such as company names, personal names, and product names in the text through the NER model, and then identify the relationships between these entities through the RE model, such as "Company A invests in Company B" and "Zhang San resigns as CEO of Company C". Then, the EE model can identify key financial events, such as "equity pledge" and "involvement in litigation", thereby obtaining corresponding triples, such as (Company A, investment, Company B), (Zhang San, resigns, Company C), (Company B, equity pledge, completed), etc.

[0096] In this embodiment of the application, the first device may also determine the triples corresponding to the fused data in other ways, which are not limited here.

[0097] In this possible implementation, the accuracy of entity and relation extraction is significantly improved through the collaborative work of deep learning models and rule engines, thereby enhancing the accuracy of entities and relations in the knowledge graph.

[0098] In one possible implementation, when the electronic device generates the corresponding triplet based on the entity recognition result, it can use a rule engine to correct the entity recognition result based on the logical rule library to obtain the first triplet.

[0099] In this embodiment of the application, the knowledge graph may include a logical rule base of knowledge in the domain. The rule engine can correct the entity recognition results based on the logical rule base of knowledge in the domain to obtain more accurate triples.

[0100] For example, in the news article "Zhang San resigns as CEO of Company C", the first device identifies the news and obtains entity recognition results: "Zhang San" (person's name), "Company C" (company name), and "resigned". Subsequently, the rule engine corrects the model output based on financial domain knowledge (such as job standardization rules), for example, converting "CEO position" into the standardized label "senior management change".

[0101] In one possible implementation, the graph database corresponding to this knowledge graph is a distributed graph database.

[0102] In this possible implementation, a distributed graph database is used to reduce the data processing pressure on a single node and improve storage and query efficiency in large-scale data scenarios.

[0103] S204. Electronic devices perform risk analysis on the updated knowledge graph using graph computing algorithms to obtain visualized risk analysis results, which include risk transmission paths.

[0104] For example, electronic devices can use a multi-hop association risk discovery algorithm to perform risk analysis on the updated knowledge graph, obtaining the risk transmission path: Company A -- (with supply chain relationship) --> Company C -- (involved) --> major environmental litigation -- (potentially) --> production stoppage and fines. Therefore, it is understandable that the risk of Company C may be transmitted to Company A through the supply chain, affecting the stability of its future production and operation.

[0105] In this embodiment of the application, the risk transmission path refers to the path by which risk spreads and is transmitted in the financial network along the connection paths between entities (such as industrial chains, guarantee chains, equity chains, or capital chains).

[0106] In one possible implementation, the electronic device can also calculate the risk transmission score based on information such as the length of the risk transmission path and the strength of the relationship. That is, the risk analysis result also includes the risk transmission score, which is not specified here.

[0107] In one possible implementation, when an electronic device performs risk analysis on an updated knowledge graph using a graph computation algorithm, it can first determine the corresponding first graph computation algorithm and first graph computation parameters based on the risk scenario type; and determine the corresponding first graph computation strategy based on the specifications of the knowledge graph; and then perform risk analysis on the knowledge graph using the first graph computation algorithm based on the first graph computation parameters and the first graph computation strategy.

[0108] In one possible implementation, the electronic device determines the corresponding first graph calculation parameters based on the risk scenario type. For different risk analysis scenarios, the risk categories to be analyzed are different, and different algorithms and calculation parameters can be set accordingly to obtain more accurate analysis results.

[0109] For example, the electronic device may include a graph computing algorithm library, which includes various graph computing algorithms such as path discovery algorithms (e.g., shortest path algorithms), community discovery algorithms, centrality algorithms, and custom risk inference rule engines. For example, in a money transfer scenario, the electronic device can call a path discovery algorithm to find the funding link; for example, in a credit approval scenario, the electronic device can find the association path between the target company and known risky entities (e.g., defaulters, companies involved in litigation).

[0110] For example, electronic devices can be configured with different calculation parameters for different risk analysis scenarios. For instance, in a credit approval scenario, the risk scenario type could be "supply chain risk," and the corresponding path hop threshold could be 4. In a money transfer scenario, the risk scenario type could be "abnormal transaction," and the corresponding path hop threshold could be 3.

[0111] This possible implementation method determines different graph computation algorithms, parameters, and / or strategies for different risk scenario types and knowledge graph specifications. This improves the adaptability of the risk analysis method to different risk scenario types and knowledge graph specifications, enhances the applicability of financial risk analysis, allows it to be applied to more complex situations, strengthens financial risk analysis capabilities, and improves the accuracy of financial risk analysis results.

[0112] In one possible implementation, the electronic device determines the corresponding first graph computation strategy based on the specifications of the knowledge graph. The electronic device can also determine the corresponding graph computation algorithm based on the specifications of the knowledge graph. For example, the electronic device can determine the corresponding graph computation algorithm based on the specifications of the knowledge graph, using an exact algorithm for smaller knowledge graphs and an approximate algorithm for larger knowledge graphs.

[0113] In this embodiment of the application, the specifications of the knowledge graph can be information indicating the size and complexity of the knowledge graph. In addition, it can also be other information, such as the number of distributed nodes, etc., which are not limited here.

[0114] In one possible implementation, when performing risk analysis, electronic devices can obtain not only relevant risk results but also risk data sources. That is, the risk analysis results can also include risk data sources, which are the original data (such as news report links, announcement screenshots, etc.) on which the risk analysis results are based.

[0115] For example, electronic devices can determine the risk data source corresponding to a risk transmission path. For instance, an electronic device can use a causal inference engine, based on a causal rule base, to determine the risk data source corresponding to a risk transmission path. This causal inference engine is a reasoning system based on logical rules and probabilistic graphical models, used to analyze the causal logic of risk transmission. For example, the causal rule base can define rules such as: "Environmental litigation may lead to enterprise shutdown → shutdown affects supply chain stability."

[0116] This causal reasoning engine can identify causal relationships in risk transmission paths. Therefore, after the graph computing algorithm analyzes the risk transmission path, the causal reasoning engine can determine the risk data source based on the attribute information of each node in the risk transmission path and the corresponding rules in the causal rule base. For example, it can analyze the causal logic of the path "environmental litigation → enterprise shutdown → supply chain disruption" and then determine the corresponding raw data as the risk data source.

[0117] In this embodiment of the application, after the electronic device performs financial risk analysis and obtains the risk analysis results, it can display the risk analysis results to the bank's risk control personnel. Therefore, optionally, the method 200 may include the following steps:

[0118] S205, Electronic devices display risk analysis results.

[0119] In one possible implementation, the electronic device may include a display module that can display the results of the risk analysis.

[0120] For example, the display module can display a web-based front-end interactive interface. The electronic device can present the results of the analysis and inference module to risk control personnel in a graphical and interactive manner.

[0121] In one possible implementation, the electronic device can display the risk analysis results corresponding to the first entity, which includes the risk transmission path corresponding to the first entity displayed in the form of text or images. For example, the risk transmission path "Company A → Company C → Environmental Litigation" can be highlighted.

[0122] In one possible implementation, the electronic device can display the risk data source corresponding to the risk transmission path. This risk data source is the original data on which the risk analysis results are based, such as news report links, announcement screenshots, etc.

[0123] For example, when an electronic device displays a risk transmission path, it can also display the corresponding risk data source and indicate the relationship between the two: the original data is the risk data source for this risk transmission path.

[0124] In this possible implementation method, risk control personnel can understand the causal logic of "environmental litigation → enterprise shutdown → supply chain disruption" through a visual interface and verify the rationality of the conclusions. By binding causal chain analysis with the original data source, the interpretability of risk conclusions is significantly improved. This helps risk control personnel understand the internal logic of risk transmission, rather than relying solely on the surface information of the correlation path, thereby improving the scientific and compliant nature of decision-making.

[0125] In one possible implementation, the electronic device can display the risk score corresponding to the first entity.

[0126] In addition, the risk analysis results displayed on electronic devices may include other information, which is not limited here.

[0127] In this embodiment of the application, the following technical effects are achieved through data fusion of unstructured and structured data, dynamic knowledge graph construction and updating, and graph analysis reasoning:

[0128] 1. By collecting external unstructured data in real time (such as news and policy announcements) and using the fused data as risk analysis data, the system can identify potential signals before risk events occur, improving the real-time nature and accuracy of financial risk identification.

[0129] 2. Graph computing algorithms can be used to mine complex relationships with multiple layers (such as equity chains, guarantee chains, capital chains, or industrial chains) and identify hidden, deep-seated associated risks and risk transmission paths.

[0130] 3. The final risk analysis results, including the risk transmission path and the corresponding risk data source, are presented to users in a visual manner, which enhances the interpretability of the risk analysis results and enables risk control personnel to verify the rationality of the conclusions.

[0131] Figure 4 This is a schematic diagram of the structure of a financial risk analysis device provided in an embodiment of this application, as shown below. Figure 4 As shown, the financial risk analysis device 400 provided in this embodiment includes:

[0132] The data fusion module 401 is used to perform data fusion processing on unstructured data and structured data to obtain fused data;

[0133] Update module 402 is used to update the knowledge graph based on fused data;

[0134] The risk analysis module 403 is used to perform risk analysis on the updated knowledge graph using graph computing algorithms to obtain visualized risk analysis results, including risk transmission paths.

[0135] In one possible implementation, the data fusion module 401 is further configured to: acquire unstructured data and structured data.

[0136] In one possible implementation, the update module 402 is specifically used to: perform entity recognition and entity relation extraction on the fused data to obtain a first triplet; create a second triplet in the knowledge graph based on the first triplet, or update the attribute information of the third triplet in the knowledge graph based on the first triplet, wherein the entities and entity relations of the third triplet are the same as those of the first triplet.

[0137] In one possible implementation, the update module 402 is specifically used to: perform entity recognition and entity relationship recognition on the fused data through a deep learning model to obtain entity recognition results; and correct the entity recognition results based on the logical rule library through a rule engine to obtain the first triplet.

[0138] In one possible implementation, the risk analysis module 403 is specifically used to: determine the corresponding first graph calculation algorithm and first graph calculation parameters according to the risk scenario type; determine the corresponding first graph calculation strategy according to the specifications of the knowledge graph; and perform risk analysis on the knowledge graph using the first graph calculation algorithm based on the first graph calculation parameters and first graph calculation strategy to obtain the risk analysis results.

[0139] In one possible implementation, the data fusion module 401 is specifically used to: perform data cleaning and data format standardization on unstructured data to obtain processed unstructured data; and perform data fusion processing on the processed unstructured data and structured data to obtain fused data.

[0140] In one possible implementation, the risk analysis results also include risk data sources. Specifically, the risk analysis module 403 is used to determine the risk data sources corresponding to the risk transmission path.

[0141] In one possible implementation, the graph database corresponding to the knowledge graph is a distributed graph database.

[0142] In one possible implementation, the financial risk analysis device 400 further includes:

[0143] Receiver module 404 is used to receive the first query request;

[0144] The determination module 405 is used to determine the first graph data sub-database corresponding to the first query request through a sharding strategy. The first graph data sub-database includes the structured data corresponding to the first query request, and the multiple graph data sub-databases corresponding to the graph database include the first graph data sub-database.

[0145] The financial risk analysis device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0146] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 5 As shown, the electronic device 500 may include a memory 501 and a processor 502. Optionally, the electronic device may also include a transceiver 503, wherein the memory 501 and the processor 502 communicate; for example, the memory 501, the processor 502 and the transceiver 503 may communicate via a communication bus 504, the memory 501 is used to store a computer program, and the processor 502 executes the computer program to implement the method of the above embodiments.

[0147] Optionally, the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps in the method embodiments disclosed in this application can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0148] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the methods in any of the above method embodiments.

[0149] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the methods in any of the above method embodiments.

[0150] All or part of the steps in the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a readable memory. When the program is executed, it performs the steps of the above method embodiments; and the aforementioned memory (storage medium) includes: read-only memory (ROM), RAM, flash memory, hard disk, solid-state drive, magnetic tape, floppy disk, optical disk, and any combination thereof.

[0151] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processing unit of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processing unit of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0152] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0153] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0154] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

[0155] In this application, the term "comprising" and its variations can refer to non-limiting inclusion; the term "or" and its variations can refer to "and / or". The terms "first", "second", etc., in this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. In this application, "multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0156] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0157] It should be further noted that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0158] It should be understood that the above-described device embodiments are merely illustrative, and the device of this application can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units, modules, or components may be combined, or integrated into another system, or some features may be ignored or not executed.

[0159] Furthermore, unless otherwise specified, the functional units / modules in the various embodiments of this application can be integrated into one unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated together. The integrated units / modules described above can be implemented in hardware or as software program modules.

[0160] When integrated units / modules are implemented in hardware, the hardware can be digital circuits, analog circuits, etc. The physical implementation of the hardware structure includes, but is not limited to, transistors, memristors, etc. Unless otherwise specified, the processor can be any suitable hardware processor, such as a CPU, GPU, FPGA, DSP, and ASIC, etc. Unless otherwise specified, the storage unit can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc.

[0161] If the integrated unit / module is implemented as a software program module and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0162] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0163] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0164] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A financial risk analysis method, characterized in that, The method includes: Unstructured and structured data are fused together to obtain fused data. The knowledge graph is updated based on the fused data; Risk analysis is performed on the updated knowledge graph using graph computing algorithms to obtain visualized risk analysis results, which include risk transmission paths.

2. The method according to claim 1, characterized in that, The updating of the knowledge graph based on the fused data includes: Entity recognition and entity relationship extraction are performed on the fused data to obtain the first triplet; Create a second triplet in the knowledge graph based on the first triplet, or... The attribute information of the third triple in the knowledge graph is updated based on the first triple, and the entities and entity relationships of the third triple are the same as those of the first triple.

3. The method according to claim 2, characterized in that, The step of performing entity recognition and entity relation extraction on the fused data to obtain the first triplet includes: The entity recognition results are obtained by performing entity recognition and entity relationship recognition on the fused data using a deep learning model. The first triplet is obtained by correcting the entity recognition result based on the logical rule base using a rule engine.

4. The method according to claim 3, characterized in that, The risk transmission path analysis is performed on the updated knowledge graph using a graph computation algorithm to obtain risk analysis results, including: Determine the corresponding first graph calculation algorithm and first graph calculation parameters based on the type of risk scenario; The corresponding first graph calculation strategy is determined based on the specifications of the knowledge graph; Based on the calculation parameters and strategy of the first graph, a risk analysis is performed on the updated knowledge graph using the first graph calculation algorithm to obtain the risk analysis results.

5. The method according to claim 4, characterized in that, The process of fusing unstructured and structured data to obtain fused data includes: The unstructured data is cleaned and its format is standardized to obtain the processed unstructured data. The fused data is obtained by performing data fusion processing on the processed unstructured data and structured data.

6. The method according to claim 5, characterized in that, The risk analysis results also include risk data sources. The risk transmission path analysis of the updated knowledge graph using graph computing algorithms yields the following risk analysis results: Identify the risk data sources corresponding to the risk transmission path.

7. The method according to claim 6, characterized in that, The graph database corresponding to the knowledge graph is a distributed graph database.

8. The method according to claim 7, characterized in that, The method further includes: Receive the first query request; The first graph data sub-database corresponding to the first query request is determined by a sharding strategy. The first graph data sub-database includes the structured data corresponding to the first query request. The multiple graph data sub-databases corresponding to the graph database include the first graph data sub-database.

9. A financial risk analysis device, characterized in that, The device includes: The data fusion module is used to perform data fusion processing on unstructured and structured data to obtain fused data; The update module is used to update the knowledge graph based on the fused data; The risk analysis module is used to perform risk analysis on the updated knowledge graph using graph computing algorithms to obtain visualized risk analysis results, including risk transmission paths.

10. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1 to 8.

12. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 8.

Images