Enterprise risk identification method and device, computer device and storage medium

Abstract

The application relates to an enterprise risk identification method and device, computer equipment, a storage medium and a computer program product. The method obtains flow information of a to-be-processed bill, constructs a corresponding knowledge graph according to relationship data of nodes in the flow information, and then obtains enterprise risk factors of each target node in the knowledge graph, and determines a risk coefficient of each target node according to the knowledge graph, the enterprise risk factors of each target node and preset first risk enterprise node information. Since the embodiment fuses multi-dimensional risk characteristics when determining the risk coefficient of the enterprise, the accuracy and stability of the risk evaluation result can be improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, computer equipment, storage medium, and computer program product for identifying enterprise risks. Background Technology

[0002] With the development of computer processing technology, risk assessment is being applied more and more widely. For example, risk assessments can now be conducted on companies that issue invoices using invoice information.

[0003] In traditional technology, at least one invoice information of a first enterprise whose invoice issuance time falls within a first preset time period is obtained. This at least one invoice information includes a first type of invoice information, which is invoice information issued by the first enterprise. The resource quantities included in the first type of invoice information are accumulated, and the accumulated value is determined as the first total resource quantity. The average value of the first total resource quantity is obtained, which is the average value of the total resource quantity received by at least one second enterprise within the first preset time period. The second enterprise is a peer enterprise located in the same region as the first enterprise. The risk of the first enterprise corresponding to the invoice information is determined based on the first total resource quantity and the average value of the first total resource quantity.

[0004] However, current practices result in a lack of accuracy and stability in risk assessment results. Summary of the Invention

[0005] Therefore, it is necessary to address the problem that relying solely on the face value of a bill to compare with industry peers in the same region to determine the risk of the company corresponding to the bill leads to a lack of accuracy and stability in risk assessment results. A method, apparatus, computer equipment, storage medium, and computer program product that can accurately and stably evaluate corporate risk should be provided.

[0006] Firstly, this application provides a method for identifying enterprise risks. The method includes:

[0007] Obtain the circulation information of the bill to be processed, the circulation information including the nodes through which the bill to be processed circulates within the corresponding life cycle and the relationship data of the nodes;

[0008] Construct a corresponding knowledge graph based on the relationship data of the nodes;

[0009] Obtain the enterprise risk factors of each target node in the knowledge graph, and determine the risk coefficient of each target node based on the knowledge graph, the enterprise risk factors of each target node, and the preset first risk enterprise node information.

[0010] In one embodiment, determining the risk coefficient of each target node based on the knowledge graph, the enterprise risk factors of each target node, and preset first risk enterprise node information includes: determining a first parameter based on the preset first risk enterprise node information, wherein the first parameter characterizes whether each target node is a risk enterprise node; using the knowledge graph and the enterprise risk factors of each target node as a second parameter; and obtaining the risk coefficient of each target node using a graph neural network algorithm based on the first parameter and the second parameter.

[0011] In one embodiment, determining the risk coefficient of each target node based on the knowledge graph, the enterprise risk factors of each target node, and preset first risk enterprise node information includes: determining a first parameter based on the preset first risk enterprise node information, wherein the first parameter characterizes whether each target node is a risk enterprise node; determining the network risk factor corresponding to each target node based on the knowledge graph; using the enterprise risk factor and the network risk factor of each target node as a third parameter; and obtaining the risk coefficient of each target node using a machine learning algorithm based on the first parameter and the third parameter.

[0012] In one embodiment, determining the network risk factor corresponding to each target node based on the knowledge graph includes: obtaining the topological parameters of the target node relative to the topological relationship of each node in the knowledge graph; and using the topological parameters as the network risk factor corresponding to the target node.

[0013] In one embodiment, determining the network risk factor corresponding to each target node based on the knowledge graph includes: dividing the knowledge graph using a community detection algorithm to obtain multiple subgraphs, each subgraph including multiple interconnected nodes and corresponding node information; determining the risk factor of each subgraph based on the node information corresponding to each node in the subgraph and preset second risk enterprise node information; and using the risk factor of each subgraph as the network risk factor of the target node corresponding to the subgraph.

[0014] In one embodiment, determining the network risk factor corresponding to each target node based on the knowledge graph includes: determining a corresponding risk subgraph from the knowledge graph based on preset first risk enterprise node information, wherein the risk subgraph includes at least one risk enterprise node; obtaining the pattern features of the risk subgraph; searching the knowledge graph for a suspected risk subgraph that matches the pattern features; and determining the suspected risk subgraph as the network risk factor corresponding to the target node in the suspected risk subgraph.

[0015] In one embodiment, determining the first parameter based on preset first risk enterprise node information includes: when there is node information matching the target node in the preset first risk enterprise node information, determining the value of the first parameter to indicate that the target node is a risk enterprise node; or, when there is no node information matching the target node in the preset first risk enterprise node information, determining the value of the first parameter to indicate that the target node is not a risk enterprise node.

[0016] In one embodiment, obtaining the enterprise risk factors of each target node in the knowledge graph includes: obtaining node data of each target node in the knowledge graph; and performing natural language processing on the node data to obtain the enterprise risk factors of the corresponding target node.

[0017] Secondly, this application also provides a business risk identification device. The device includes:

[0018] The bill circulation information acquisition module is used to acquire the circulation information of the bill to be processed. The circulation information includes the nodes through which the bill to be processed circulates within its corresponding life cycle and the relationship data of the nodes.

[0019] The knowledge graph construction module is used to construct a corresponding knowledge graph based on the relationship data of the nodes.

[0020] The risk coefficient determination module is used to obtain the enterprise risk factors of each target node in the knowledge graph, and determine the risk coefficient of each target node based on the knowledge graph, the enterprise risk factors of each target node, and the preset first risk enterprise node information.

[0021] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the enterprise risk identification method described in the first aspect above.

[0022] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of the enterprise risk identification method described in the first aspect above.

[0023] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of the enterprise risk identification method described in the first aspect above.

[0024] The aforementioned enterprise risk identification method, apparatus, computer equipment, storage medium, and computer program product acquire the circulation information of the invoices to be processed, construct a corresponding knowledge graph based on the relationship data of the nodes in the circulation information, and then obtain the enterprise risk factors of each target node in the knowledge graph. Based on the knowledge graph, the enterprise risk factors of each target node, and the preset first risk enterprise node information, the risk coefficient of each target node is determined. Because this embodiment integrates multi-dimensional risk characteristics when determining the enterprise's risk coefficient, it can improve the accuracy and stability of the risk assessment results. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating a corporate risk identification method in one embodiment;

[0026] Figure 2 This is a flowchart illustrating the steps for determining the risk coefficient of a target node in one embodiment;

[0027] Figure 3 This is a flowchart illustrating the steps for determining the risk coefficient of a target node in another embodiment;

[0028] Figure 4 This is a flowchart illustrating the steps for determining network risk factors in one embodiment;

[0029] Figure 5 This is a flowchart illustrating the steps for determining network risk factors in another embodiment;

[0030] Figure 6 This is a flowchart illustrating the steps for determining network risk factors in yet another embodiment;

[0031] Figure 7A This is a schematic diagram of a risk subgraph in one embodiment;

[0032] Figure 7B This is a schematic diagram of a knowledge graph in one embodiment;

[0033] Figure 7C1 This is a schematic diagram of a potentially risky subgraph within a knowledge graph.

[0034] Figure 7C2 This is a schematic diagram of another potentially risky subgraph within a knowledge graph.

[0035] Figure 8 This is a flowchart illustrating the steps for determining enterprise risk factors in one embodiment;

[0036] Figure 9 This is a structural block diagram of an enterprise risk identification device in one embodiment;

[0037] Figure 10This is an internal structural diagram of a computer device in one embodiment;

[0038] Figure 11 This is a diagram of the internal structure of a computer device in another embodiment. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0040] In one embodiment, such as Figure 1 As shown, a method for enterprise risk identification is provided. This embodiment illustrates the application of this method to an identification device. It is understood that this method can also be applied to a server, or to a system including both an identification device and a server, and is implemented through the interaction between the identification device and the server. Specifically, the identification device can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices, and the server can be a standalone server or a server cluster consisting of multiple servers.

[0041] In this embodiment, the method may include the following steps:

[0042] Step 102: Obtain the transfer information of the invoices to be processed.

[0043] Among them, a pending negotiable instrument is a document issued by the drawer in accordance with the law, obligating the drawer or instructing others to pay a certain amount of money to the payee or holder. The circulation information includes the nodes in the pending negotiable instrument's circulation within its corresponding lifecycle, as well as the relationship data between these nodes. Specifically, the lifecycle of a pending negotiable instrument includes, but is not limited to, issuance, acceptance, transfer and endorsement, discounting, rediscounting, rediscounting again, maturity payment, and circulation processes such as pledging, guaranteeing, and payment.

[0044] The nodes in the circulation refer to the individuals, enterprises, and third-party regulatory agencies that the bill passes through during its corresponding lifecycle. The relational data of these nodes includes, on the one hand, the circulation relationships between nodes; for example, if a bill flows from node A to node B, the relational data includes the directed relationship from node A to node B. On the other hand, it also includes the related relationships analyzed based on the relevant data of each node. For example, for node A in the bill circulation, further relevant data about node A can be obtained, including but not limited to node A's basic information, credit information, financial statements, cash flow, business registration information, and credit information. Further analysis of this information can reveal various relationships related to node A, such as the multi-dimensional relationships between node A and other individuals, enterprises, and third-party regulatory agencies.

[0045] Specifically, in this embodiment, when it is necessary to identify the corporate risks involved in the bills, the identification device first needs to obtain the circulation information of the bills to be processed, and then perform risk identification based on subsequent steps.

[0046] Step 104: Construct the corresponding knowledge graph based on the relationship data of the nodes.

[0047] Knowledge graphs are a modern theory that combines theories and methods from applied mathematics, computer graphics, information visualization technology, and information science with methods such as bibliometric citation analysis and co-occurrence analysis. They use visualized graphs to vividly display the corresponding core structure, development history, cutting-edge fields, and overall knowledge architecture to achieve multi-dimensional integration.

[0048] In this embodiment, the identification device can construct a corresponding knowledge graph based on the relationship data of the nodes obtained above. Specifically, the knowledge graph includes nodes and edges. For example, if a ticket flows from node A to node B, the relationship data includes the directed relationship from node A to node B. In the process of generating the knowledge graph, the corresponding nodes in the graph are generated based on node A and node B, and the edges in the graph are generated based on the directed relationship from node A to node B.

[0049] Based on this, the identification device can construct a corresponding visualized knowledge graph by analyzing the relationship data of the nodes through which the ticket flows within its corresponding lifecycle, obtained from the circulation information of the ticket to be processed in the above steps.

[0050] Step 106: Obtain the enterprise risk factors for each target node in the knowledge graph and determine the risk coefficient of each target node.

[0051] In this context, the target node refers to a node whose node attribute is an enterprise. Enterprise risk factors are the potential factors contributing to enterprise risk events, representing both indirect and intrinsic causes of losses. The risk coefficient is a numerical value specifically indicating the degree of risk.

[0052] In this embodiment, the identification device acquires the enterprise risk factors of each target node in the knowledge graph, and determines the risk coefficient of each target node based on the knowledge graph, the enterprise risk factors of each target node, and the preset first risk enterprise node information. The preset first risk enterprise node information refers to pre-set, verified information about risky enterprise nodes (such as enterprise nodes with bill discounting activities), which may include, for example, at least the enterprise identifier of each risky enterprise node.

[0053] In the aforementioned enterprise risk identification method, the circulation information of the invoices to be processed is obtained, and a corresponding knowledge graph is constructed based on the relationship data of the nodes in the circulation information. Then, the enterprise risk factors of each target node in the knowledge graph are obtained. Based on the knowledge graph, the enterprise risk factors of each target node, and the preset information of the first risk enterprise node, the risk coefficient of each target node is determined. Because this embodiment integrates multi-dimensional risk characteristics when determining the enterprise's risk coefficient, it can improve the accuracy and stability of the risk assessment results.

[0054] In one embodiment, such as Figure 2 As shown, the risk coefficient of each target node is determined based on the knowledge graph, the enterprise risk factors of each target node, and the preset information of the first risk enterprise node. Specifically, it may also include:

[0055] Step 202: Determine the first parameter based on the preset first risk enterprise node information.

[0056] The first parameter characterizes whether each target node is a risky enterprise node. Since the preset first risky enterprise node information refers to pre-defined, verified information about enterprise nodes with bill discounting activities, in this implementation, the identification device can determine whether a target node is a bill discounting enterprise node based on the preset first risky enterprise node information.

[0057] Specifically, when there is node information matching the target node in the preset first risk enterprise node information, the value of the first parameter of the target node is determined to indicate that the target node is a bill discounting enterprise node; and when there is no node information matching the target node in the preset first risk enterprise node information, the value of the first parameter of the target node is determined to indicate that the target node is not a bill discounting enterprise node.

[0058] Step 204: Use the knowledge graph and the enterprise risk factors of each target node as the second parameter.

[0059] In this embodiment, after determining the first parameter, the identification device can also use the knowledge graph and the enterprise risk factors of each target node as the second parameter. Specifically, both the first and second parameters are relevant bases for determining the risk coefficient of the target node.

[0060] Step 206: Based on the first parameter and the second parameter, the risk coefficient of each target node is obtained using a graph neural network algorithm.

[0061] The graph neural network algorithms include, but are not limited to, those based on GCN (Graph Convolutional Networks) and GraphSAGE (Graph Simple and AggeGatE). Specifically, by using the second parameter as variable X and the first parameter as variable Y, the graph neural network algorithm is directly trained to output the risk coefficient of each target node in the knowledge graph. Generally, a higher risk coefficient indicates a higher risk for the corresponding target node, while a lower risk coefficient indicates a lower risk for the corresponding target node.

[0062] Because this embodiment considers not only the enterprise risk factors of the target node itself, but also the corresponding knowledge graph and the preset first risk enterprise node information when assessing the risk of the target node, and uses a graph neural network algorithm to quantify and assess the corresponding risk coefficient, it can improve the accuracy and stability of the risk assessment results, and the assessment results are more intuitive.

[0063] In one embodiment, such as Figure 3 As shown, the risk coefficient of each target node is determined based on the knowledge graph, the enterprise risk factors of each target node, and the preset information of the first risk enterprise node. Specifically, it may also include:

[0064] Step 302: Determine the first parameter based on the preset first risk enterprise node information.

[0065] For details, please refer to the description of step 202 above; this embodiment will not repeat it further.

[0066] Step 304: Based on the knowledge graph, determine the network risk factors corresponding to each target node.

[0067] Among them, network risk factors are the potential factors that determine the location of a target node within the network structure based on the knowledge graph, leading to a risk event (such as a bill discounting event). Specifically, the identification device can determine the network risk factors corresponding to each target node based on the knowledge graph.

[0068] Step 306: Use the enterprise risk factor and network risk factor of each target node as the third parameter.

[0069] In this embodiment, after determining the first parameter, the identification device can also use the enterprise risk factor and network risk factor of each target node as the third parameter. Specifically, both the first and third parameters are relevant bases for determining the risk coefficient of the target node.

[0070] Step 308: Based on the first parameter and the third parameter, a machine learning algorithm is used to obtain the risk coefficient of each target node.

[0071] Machine learning algorithms are a multidisciplinary field and the core of artificial intelligence, involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory, and many other disciplines. They specifically study how computers can simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. In this embodiment, XGBoost (an optimized distributed gradient boosting library that implements machine learning algorithms within the Gradient Boosting framework) is used as an example. By using the third parameter as the X variable and the first parameter as the Y variable, the model is trained using XGBoost to output the risk coefficient of each target node in the knowledge graph. Generally, a higher risk coefficient indicates a higher risk for the corresponding target node, while a lower risk coefficient indicates a lower risk. In this embodiment, the enterprise risk factor and network risk factor of each target node can also be output as risk clues for reference during subsequent investigations.

[0072] Because this embodiment considers not only the enterprise risk factor of the target node itself, but also the corresponding network risk factor and the preset first risk enterprise node information when assessing the risk of the target node, and uses machine learning algorithms to quantify and evaluate the corresponding risk coefficient, it can improve the accuracy and stability of the risk assessment results, and the assessment results are more intuitive.

[0073] In one embodiment, such as Figure 4 As shown, in step 304, the network risk factors corresponding to each target node are determined based on the knowledge graph, which may specifically include:

[0074] Step 402: Based on the topological relationships of each node in the knowledge graph, obtain the topological parameters of the target node relative to the topological relationships.

[0075] Topological relationships refer to the interrelationships between spatial data that satisfy the principles of topological geometry; that is, the adjacency, association, inclusion, and connectivity relationships between entities represented by nodes and edges in a knowledge graph. Topological parameters are relevant parameters determined based on topological relationships, including but not limited to parameters such as Degree, Betweenness, First-order Influence, and Second-order Influence of each target node. Degree refers to the number of edges associated with the target node, also known as the degree of association. Betweenness includes edge betweenness and node betweenness. Node betweenness is the proportion of all shortest paths in the network that pass through the target node; edge betweenness is the proportion of all shortest paths in the network that pass through that edge. First-order influence can be determined by the number of edges directly associated with the target node, while second-order influence can be determined by the number of edges associated with other nodes directly connected to the target node. These topology parameters can measure the influence, importance, or information transmission ability of each target node in the network. The higher the index, the greater the possibility of risk.

[0076] Specifically, in this embodiment, the identification device can obtain the topological parameters of the target node relative to the topological relationship of each node in the knowledge graph.

[0077] Step 404: Use the topology parameters as the network risk factors corresponding to the target node.

[0078] Since the aforementioned topology parameters can measure the influence, importance, or information transmission capability of each target node in the network, in this embodiment, the obtained topology parameters can be directly used as the network risk factor corresponding to the target node. Furthermore, by fusing network risk factors to assess the risk coefficient of the target node, the stability of the assessment can be improved.

[0079] In one embodiment, such as Figure 5 As shown, in step 304, the network risk factors corresponding to each target node are determined based on the knowledge graph, which may specifically include:

[0080] Step 502: The knowledge graph is divided using a community detection algorithm to obtain multiple subgraphs.

[0081] Community detection algorithms include, but are not limited to, Louvain (a modularity-based community detection algorithm), LPA (Label Propagation Algorithm), and connected component algorithms. In this embodiment, community detection algorithms can group nodes with high inter-relationships in the knowledge graph together, resulting in multiple subgraphs. Each subgraph can include multiple inter-relationship nodes and their corresponding node information. Specifically, node information may include the amount of tickets circulated corresponding to that node, or the transaction amount corresponding to that node. In this embodiment, the identification device can use community detection algorithms to divide the knowledge graph, thereby obtaining multiple subgraphs.

[0082] Step 504: Determine the network risk factor of the target node corresponding to the subgraph based on the node information corresponding to each node in the subgraph and the preset second risk enterprise node information.

[0083] The preset second-risk enterprise node information may include pre-set suspicious enterprise nodes for bill discounting and their corresponding node information. Specifically, the node information may include the enterprise identifier of the suspicious enterprise node for bill discounting, and the amount of bills or transactions involved in each suspicious enterprise node for bill discounting. In this embodiment, the identification device can determine the risk factor of each subgraph based on the node information corresponding to each node in the subgraph and the preset second-risk enterprise node information, and use the risk factor of each subgraph as the network risk factor of the target node corresponding to the subgraph.

[0084] Specifically, the risk factors of a subgraph can be the proportion of suspicious bill discounting enterprise nodes in that subgraph and the proportion of bill amount or transaction amount corresponding to those suspicious bill discounting enterprise nodes. For example, if a subgraph has 5 nodes, and based on preset second-risk enterprise node information, it can be determined that there are 2 suspicious bill discounting enterprise nodes among them, then the proportion of suspicious bill discounting enterprise nodes in the corresponding subgraph is (2 / 5), or 40%. Similarly, based on the sum of bill amount or transaction amount of all nodes in the subgraph, and the bill amount or transaction amount involved in the suspicious bill discounting enterprise nodes in the subgraph, the proportion of bill amount or transaction amount corresponding to the suspicious bill discounting enterprise nodes in the corresponding subgraph can be calculated. Based on this, the risk factors of each subgraph can be obtained, and further, the risk factors of each subgraph can be used as the network risk factors of the target nodes in that subgraph.

[0085] This embodiment uses a community discovery algorithm to divide the knowledge graph into multiple subgraphs. Based on the node information corresponding to each node in the subgraph and the preset second risk enterprise node information, the network risk factor of the target node corresponding to the subgraph is determined. Then, the risk coefficient of the target node can be evaluated by fusing the network risk factors to improve the stability of the evaluation.

[0086] In one embodiment, such as Figure 6 As shown, in step 304, the network risk factors corresponding to each target node are determined based on the knowledge graph, which may specifically include:

[0087] Step 602: Based on the preset first risk enterprise node information, determine the corresponding risk subgraph from the knowledge graph.

[0088] The risk subgraph includes at least one risky enterprise node (i.e., a verified bill discounting enterprise node). Since the information of the first risky enterprise node is pre-set information about verified risky enterprise nodes (i.e., enterprise nodes with bill discounting activities), the identification device can determine the corresponding risk subgraph from the knowledge graph based on the pre-set information of the first risky enterprise node.

[0089] Specifically, the identification device divides nodes in the knowledge graph that have a high degree of correlation with risk enterprise nodes in the preset first risk enterprise node information together to obtain a divided risk subgraph. Each risk subgraph includes at least one risk enterprise node.

[0090] Step 604: Obtain the pattern features of the risk subgraph and search for suspected risk subgraphs that match the pattern features in the knowledge graph.

[0091] Pattern features are the basis for recognition in pattern recognition. For example, in image pattern recognition, one of the various measurements of a pattern or a mathematical transformation of these measurements, such as grayscale, texture, and shape, can be called a pattern feature. A suspected risk subgraph can be a subgraph in a knowledge graph that has the same pattern features as a risk subgraph.

[0092] In this embodiment, the pattern feature can be the graph structure corresponding to the risk subgraph, that is, the relationship structure between nodes in the risk subgraph. Specifically, the identification device obtains the pattern feature of the risk subgraph and uses a pattern matching algorithm to search for suspected risk subgraphs that match the pattern feature in the knowledge graph.

[0093] Step 606: Identify the suspected risk subgraph as the network risk factor corresponding to the target node in the suspected risk subgraph.

[0094] Specifically, the identification device identifies the suspected risk subgraphs obtained above as network risk factors corresponding to the target nodes in the suspected risk subgraphs.

[0095] For example, if the identification device determines the risk subgraph from the knowledge graph based on the preset first risk enterprise node information, such as... Figure 7A As shown, the pattern features of the risk subgraph are then further obtained (from...). Figure 7A It can be seen that its pattern is characterized by nodes being connected sequentially to form a closed loop, i.e., node A points to node C, node C points to node G, and node G points back to node A, thus forming a closed loop. If the knowledge graph is like this... Figure 7B As shown, pattern matching algorithms can be used to analyze knowledge graphs. Figure 7B Searching for and such Figure 7A The pattern features of the risk subgraph match the suspected risk subgraph, i.e., in the knowledge graph. Figure 7B Search for the risk subgraph in the middle. Figure 7A Other subgraphs with the same graph structure are identified, and these subgraphs are considered potential risk subgraphs. For example, if subgraphs 7C1 and 7C2 are found, then... Figure 7C1 This can serve as a network risk factor corresponding to the target node in the graph; correspondingly, the sub-nodes... Figure 7C2 It can also serve as a network risk factor for the target node in the corresponding graph. Furthermore, by fusing network risk factors, the risk coefficient of the target node can be assessed to improve the stability of the assessment.

[0096] In one embodiment, such as Figure 8 As shown, in step 106, the enterprise risk factors of each target node in the knowledge graph are obtained, which may specifically include:

[0097] Step 802: Obtain the node data of each target node in the knowledge graph.

[0098] The node data includes, but is not limited to, data on judicial litigation, administrative penalties, public opinion, business registration, financial matters, and credit defaults corresponding to the target node. In this embodiment, the identification device acquires the node data of each target node in the knowledge graph and determines the corresponding enterprise risk factors based on subsequent steps.

[0099] Step 804: Perform natural language processing on the node data to obtain the enterprise risk factor for the corresponding target node.

[0100] Natural Language Processing (NLP) is a science that integrates linguistics, computer science, and mathematics, enabling effective communication between humans and computers using natural language. In this embodiment, the identification device performs NLP on the node data of the aforementioned target node to obtain the corresponding enterprise risk factors. Specifically, the identification device uses NLP technology to analyze the text information in the judicial litigation data, administrative penalty data, public opinion data, business registration data, financial data, and credit default data of the corresponding target node, and transforms it into risk factors such as judicial risk level, administrative risk level, public opinion risk level, business registration risk level, financial risk level, and credit risk level. These risk factors are then used as the enterprise risk factors for the corresponding target node. Furthermore, by integrating these enterprise risk factors, the risk coefficient of the target node can be assessed, thereby improving the stability and accuracy of the assessment.

[0101] It should be understood that although the steps in the flowcharts of the embodiments described above 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 flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0102] Based on the same inventive concept, this application also provides an enterprise risk identification device for implementing the enterprise risk identification method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more enterprise risk identification device embodiments provided below can be found in the limitations of the enterprise risk identification method described above, and will not be repeated here.

[0103] In one embodiment, such as Figure 9 As shown, a corporate risk identification device is provided, including: a bill circulation information acquisition module 902, a knowledge graph construction module 904, and a risk coefficient determination module 906, wherein:

[0104] The bill circulation information acquisition module 902 is used to acquire the circulation information of the bill to be processed, the circulation information including the nodes through which the bill to be processed circulates within the corresponding life cycle and the relationship data of the nodes;

[0105] The knowledge graph construction module 904 is used to construct a corresponding knowledge graph based on the relationship data of the nodes.

[0106] The risk coefficient determination module 906 is used to obtain the enterprise risk factors of each target node in the knowledge graph, and determine the risk coefficient of each target node based on the knowledge graph, the enterprise risk factors of each target node, and the preset first risk enterprise node information.

[0107] In one embodiment, the risk coefficient determination module includes: a first parameter determination unit, configured to determine a first parameter based on preset first risk enterprise node information, wherein the first parameter characterizes whether each target node is a risk enterprise node; a second parameter determination unit, configured to use the knowledge graph and the enterprise risk factor of each target node as the second parameter; and a risk coefficient acquisition unit, configured to acquire the risk coefficient of each target node using a graph neural network algorithm based on the first parameter and the second parameter.

[0108] In one embodiment, the risk coefficient determination module includes: a first parameter determination unit, configured to determine a first parameter based on preset first risk enterprise node information, wherein the first parameter characterizes whether each target node is a risk enterprise node; a third parameter determination unit, configured to determine the network risk factor corresponding to each target node based on the knowledge graph, and use the enterprise risk factor and the network risk factor of each target node as the third parameter; and a risk coefficient acquisition unit, configured to acquire the risk coefficient of each target node using a machine learning algorithm based on the first parameter and the third parameter.

[0109] In one embodiment, the third parameter determination unit is further configured to: obtain the topological parameters of the target node relative to the topological relationship based on the topological relationship of each node in the knowledge graph; and use the topological parameters as the network risk factor corresponding to the target node.

[0110] In one embodiment, the third parameter determination unit is further configured to: divide the knowledge graph using a community discovery algorithm to obtain multiple subgraphs, each subgraph including multiple interconnected nodes and corresponding node information; determine the risk factor of each subgraph based on the node information corresponding to each node in the subgraph and the preset second risk enterprise node information, and use the risk factor of each subgraph as the network risk factor of the target node corresponding to the subgraph.

[0111] In one embodiment, the third parameter determination unit is further configured to: determine a corresponding risk subgraph from the knowledge graph based on preset first risk enterprise node information, wherein the risk subgraph includes at least one first risk enterprise node; obtain the pattern features of the risk subgraph, search for a suspected risk subgraph that matches the pattern features from the knowledge graph; and determine the suspected risk subgraph as the network risk factor corresponding to the target node in the suspected risk subgraph.

[0112] In one embodiment, the first parameter determining unit is further configured to: when there is node information matching the target node in the preset first risk enterprise node information, determine that the value of the first parameter indicates that the target node is a risk enterprise node; or, when there is no node information matching the target node in the preset first risk enterprise node information, determine that the value of the first parameter indicates that the target node is not a risk enterprise node.

[0113] In one embodiment, the risk coefficient determination module is further configured to: acquire node data of each target node in the knowledge graph; perform natural language processing on the node data to obtain the enterprise risk factor of the corresponding target node.

[0114] Each module in the aforementioned enterprise risk identification device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0115] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 10 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores relevant data for each node. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements an enterprise risk identification method.

[0116] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 11As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a corporate risk identification method. The display screen can be an LCD screen or an e-ink display screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0117] Those skilled in the art will understand that Figure 10 and Figure 11 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0118] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0119] Obtain the circulation information of the bill to be processed, the circulation information including the nodes through which the bill to be processed circulates within the corresponding life cycle and the relationship data of the nodes;

[0120] Construct a corresponding knowledge graph based on the relationship data of the nodes;

[0121] Obtain the enterprise risk factors of each target node in the knowledge graph, and determine the risk coefficient of each target node based on the knowledge graph, the enterprise risk factors of each target node, and the preset first risk enterprise node information.

[0122] In one embodiment, when the processor executes the computer program, it further performs the following steps: determining a first parameter based on preset first risk enterprise node information, wherein the first parameter characterizes whether each target node is a risk enterprise node; using the knowledge graph and the enterprise risk factor of each target node as a second parameter; and obtaining the risk coefficient of each target node using a graph neural network algorithm based on the first parameter and the second parameter.

[0123] In one embodiment, when the processor executes the computer program, it further performs the following steps: determining a first parameter based on preset first risk enterprise node information, wherein the first parameter characterizes whether each target node is a risk enterprise node; determining the network risk factor corresponding to each target node based on the knowledge graph; using the enterprise risk factor and the network risk factor of each target node as a third parameter; and obtaining the risk coefficient of each target node using a machine learning algorithm based on the first parameter and the third parameter.

[0124] In one embodiment, when the processor executes the computer program, it further performs the following steps: obtaining the topological parameters of the target node relative to the topological relationship of each node in the knowledge graph; and using the topological parameters as the network risk factor corresponding to the target node.

[0125] In one embodiment, when the processor executes the computer program, it further performs the following steps: dividing the knowledge graph using a community discovery algorithm to obtain multiple subgraphs, each subgraph including multiple interconnected nodes and corresponding node information; determining the risk factor of each subgraph based on the node information corresponding to each node in the subgraph and the preset second risk enterprise node information; and using the risk factor of each subgraph as the network risk factor of the target node corresponding to the subgraph.

[0126] In one embodiment, when the processor executes the computer program, it further performs the following steps: determining a corresponding risk subgraph from the knowledge graph based on preset first risk enterprise node information, wherein the risk subgraph includes at least one risk enterprise node; obtaining the pattern features of the risk subgraph, searching the knowledge graph for a suspected risk subgraph that matches the pattern features; and determining the suspected risk subgraph as the network risk factor corresponding to the target node in the suspected risk subgraph.

[0127] In one embodiment, when the processor executes the computer program, it further implements the following steps: when there is node information matching the target node in the preset first risk enterprise node information, it determines that the value of the first parameter indicates that the target node is a risk enterprise node; or, when there is no node information matching the target node in the preset first risk enterprise node information, it determines that the value of the first parameter indicates that the target node is not a risk enterprise node.

[0128] In one embodiment, when the processor executes the computer program, it further performs the following steps: acquiring node data of each target node in the knowledge graph; performing natural language processing on the node data to obtain the enterprise risk factor of the corresponding target node.

[0129] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0130] Obtain the circulation information of the bill to be processed, the circulation information including the nodes through which the bill to be processed circulates within the corresponding life cycle and the relationship data of the nodes;

[0131] Construct a corresponding knowledge graph based on the relationship data of the nodes;

[0132] Obtain the enterprise risk factors of each target node in the knowledge graph, and determine the risk coefficient of each target node based on the knowledge graph, the enterprise risk factors of each target node, and the preset first risk enterprise node information.

[0133] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: determining a first parameter based on preset first risk enterprise node information, wherein the first parameter characterizes whether each target node is a risk enterprise node; using the knowledge graph and the enterprise risk factor of each target node as a second parameter; and obtaining the risk coefficient of each target node using a graph neural network algorithm based on the first parameter and the second parameter.

[0134] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: determining a first parameter based on preset first risk enterprise node information, wherein the first parameter characterizes whether each target node is a risk enterprise node; determining the network risk factor corresponding to each target node based on the knowledge graph; using the enterprise risk factor and the network risk factor of each target node as a third parameter; and obtaining the risk coefficient of each target node using a machine learning algorithm based on the first parameter and the third parameter.

[0135] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: obtaining the topological parameters of the target node relative to the topological relationships of each node in the knowledge graph; and using the topological parameters as the network risk factor corresponding to the target node.

[0136] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: dividing the knowledge graph using a community discovery algorithm to obtain multiple subgraphs, each subgraph including multiple interconnected nodes and corresponding node information; determining the risk factor of each subgraph based on the node information corresponding to each node in the subgraph and the preset second risk enterprise node information; and using the risk factor of each subgraph as the network risk factor of the target node corresponding to the subgraph.

[0137] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: determining a corresponding risk subgraph from the knowledge graph based on preset first risk enterprise node information, wherein the risk subgraph includes at least one risk enterprise node; obtaining the pattern features of the risk subgraph, searching the knowledge graph for a suspected risk subgraph that matches the pattern features; and determining the suspected risk subgraph as the network risk factor corresponding to the target node in the suspected risk subgraph.

[0138] In one embodiment, when the computer program is executed by the processor, it further implements the following steps: when there is node information matching the target node in the preset first risk enterprise node information, it determines that the value of the first parameter indicates that the target node is a risk enterprise node; or, when there is no node information matching the target node in the preset first risk enterprise node information, it determines that the value of the first parameter indicates that the target node is not a risk enterprise node.

[0139] In one embodiment, when the computer program is executed by a processor, it further performs the following steps: acquiring node data of each target node in the knowledge graph; performing natural language processing on the node data to obtain the enterprise risk factor of the corresponding target node.

[0140] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps:

[0141] Obtain the circulation information of the bill to be processed, the circulation information including the nodes through which the bill to be processed circulates within the corresponding life cycle and the relationship data of the nodes;

[0142] Construct a corresponding knowledge graph based on the relationship data of the nodes;

[0143] Obtain the enterprise risk factors of each target node in the knowledge graph, and determine the risk coefficient of each target node based on the knowledge graph, the enterprise risk factors of each target node, and the preset first risk enterprise node information.

[0144] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: determining a first parameter based on preset first risk enterprise node information, wherein the first parameter characterizes whether each target node is a risk enterprise node; using the knowledge graph and the enterprise risk factor of each target node as a second parameter; and obtaining the risk coefficient of each target node using a graph neural network algorithm based on the first parameter and the second parameter.

[0145] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: determining a first parameter based on preset first risk enterprise node information, wherein the first parameter characterizes whether each target node is a risk enterprise node; determining the network risk factor corresponding to each target node based on the knowledge graph; using the enterprise risk factor and the network risk factor of each target node as a third parameter; and obtaining the risk coefficient of each target node using a machine learning algorithm based on the first parameter and the third parameter.

[0146] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: obtaining the topological parameters of the target node relative to the topological relationships of each node in the knowledge graph; and using the topological parameters as the network risk factor corresponding to the target node.

[0147] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: dividing the knowledge graph using a community discovery algorithm to obtain multiple subgraphs, each subgraph including multiple interconnected nodes and corresponding node information; determining the risk factor of each subgraph based on the node information corresponding to each node in the subgraph and the preset second risk enterprise node information; and using the risk factor of each subgraph as the network risk factor of the target node corresponding to the subgraph.

[0148] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: determining a corresponding risk subgraph from the knowledge graph based on preset first risk enterprise node information, wherein the risk subgraph includes at least one risk enterprise node; obtaining the pattern features of the risk subgraph, searching the knowledge graph for a suspected risk subgraph that matches the pattern features; and determining the suspected risk subgraph as the network risk factor corresponding to the target node in the suspected risk subgraph.

[0149] In one embodiment, when the computer program is executed by the processor, it further implements the following steps: when there is node information matching the target node in the preset first risk enterprise node information, it determines that the value of the first parameter indicates that the target node is a risk enterprise node; or, when there is no node information matching the target node in the preset first risk enterprise node information, it determines that the value of the first parameter indicates that the target node is not a risk enterprise node.

[0150] In one embodiment, when the computer program is executed by a processor, it further performs the following steps: acquiring node data of each target node in the knowledge graph; performing natural language processing on the node data to obtain the enterprise risk factor of the corresponding target node.

[0151] 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.

[0152] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0153] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0154] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method of enterprise risk identification, characterized by, The method comprises: acquiring flow information of a to-be-processed bill, the flow information comprising nodes through which the to-be-processed bill flows in a corresponding life cycle and relationship data of the nodes; the relationship data of the nodes comprising flow relationship between the to-be-processed bill at each node and correlation relationship parsed based on relevant data of each node; constructing a corresponding knowledge graph according to the relationship data of the nodes; acquiring enterprise risk factors of each target node in the knowledge graph, and determining risk coefficients of the target nodes according to the knowledge graph, the enterprise risk factors of the target nodes, and preset first risk enterprise node information; the determination of the risk coefficients of the target nodes according to the knowledge graph, the enterprise risk factors of the target nodes, and the preset first risk enterprise node information comprises: when there is node information matching the target nodes in the preset first risk enterprise node information, determining a first parameter representing that the target nodes are risk enterprise nodes; when there is no node information matching the target nodes in the preset first risk enterprise node information, determining a first parameter representing that the target nodes are not risk enterprise nodes; taking the knowledge graph and the enterprise risk factors of the target nodes as second parameters, and acquiring the risk coefficients of the target nodes by using a graph neural network algorithm according to the first parameters and the second parameters; or determining network risk factors corresponding to each target node according to the knowledge graph, taking the enterprise risk factors of the target nodes and the network risk factors as third parameters, and acquiring the risk coefficients of the target nodes by using a machine learning algorithm according to the first parameters and the third parameters.

2. The method of claim 1, wherein, the determination of the network risk factors corresponding to each target node according to the knowledge graph comprises: acquiring a topological parameter of the target nodes relative to a topological relationship of nodes in the knowledge graph according to the topological relationship; taking the topological parameter as the network risk factor corresponding to the target nodes.

3. The method of claim 1, wherein, the determination of the network risk factors corresponding to each target node according to the knowledge graph comprises: dividing the knowledge graph by using a community discovery algorithm to obtain a plurality of divided subgraphs, each of the subgraphs comprising a plurality of mutually associated nodes and corresponding node information; determining risk factors of each subgraph according to the node information corresponding to each node in the subgraph and preset second risk enterprise node information, and taking the risk factors of each subgraph as network risk factors of target nodes corresponding to the subgraph.

4. The method of claim 1, wherein, the determination of the network risk factors corresponding to each target node according to the knowledge graph comprises: determining a corresponding risk subgraph from the knowledge graph according to preset first risk enterprise node information, the risk subgraph comprising at least one risk enterprise node; acquiring a pattern feature of the risk subgraph, and searching for a suspected risk subgraph matching the pattern feature from the knowledge graph; determining the suspected risk subgraph as network risk factors of target nodes in the suspected risk subgraph.

5. The method according to any one of claims 1 to 4, characterized in that, The enterprise risk factor of each target node in the knowledge graph is obtained, and includes the following steps: Obtaining node data of each target node in the knowledge graph; Performing natural language processing on the node data to obtain the enterprise risk factor of the corresponding target node.

6. An enterprise risk identification apparatus characterized by comprising: The device includes: A bill flow information acquisition module, configured to acquire flow information of a to-be-processed bill, the flow information including nodes through which the to-be-processed bill flows in a corresponding life cycle and relationship data of the nodes; the relationship data of the nodes including flow relationship between the to-be-processed bill and each node and an association relationship parsed based on relevant data of each node; A knowledge graph construction module, configured to construct a corresponding knowledge graph according to the relationship data of the nodes; A risk coefficient determination module, configured to obtain enterprise risk factors of each target node in the knowledge graph, and determine risk coefficients of the target nodes according to the knowledge graph, the enterprise risk factors of the target nodes, and preset first risk enterprise node information; The risk coefficient determination module is specifically configured to: When there is node information matching the target node in the preset first risk enterprise node information, a first parameter representing that the target node is a risk enterprise node is determined; or When there is no node information matching the target node in the preset first risk enterprise node information, a first parameter representing that the target node is not a risk enterprise node is determined; The knowledge graph and the enterprise risk factors of each target node are taken as a second parameter; According to the first parameter and the second parameter, a graph neural network algorithm is used to obtain the risk coefficients of the target nodes. 7.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-6 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the enterprise risk identification method in any one of claims 1 to 5.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the enterprise risk identification method in any one of claims 1 to 5.

9. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the enterprise risk identification method in any one of claims 1 to 5.

Images