Risk identification method and device for enterprise insurance, electronic equipment and storage medium

Abstract

The embodiment of the invention provides an enterprise insurance risk identification method and device, electronic equipment and a storage medium, and belongs to the technical field of artificial intelligence, and the method comprises the steps: obtaining a plurality of pieces of historical enterprise insurance policy information, a plurality of pieces of claim settlement information and a plurality of pieces of enterprise association information corresponding to a target enterprise insurance, and carrying out the association to obtain a plurality of pieces of reference information; inputting the reference information into the machine learning model, and outputting a risk assessment result; constructing a target relation graph based on the enterprise attribute information, the employee attribute information, the risk assessment result and the employment relation information; and determining abnormal enterprise information and abnormal employee information on the basis of the risk identification rule and the target relation graph, and determining the abnormal enterprise information and the abnormal employee information as target risk identification results. And the abnormal enterprise information and the abnormal employee information in the relation graph are identified in combination with the risk identification rule, so that the accuracy of the risk identification result of the enterprise insurance is improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a risk identification method and device, electronic equipment and storage medium for enterprise insurance. Background Art

[0002] When enterprise insurance risk assessment systems identify risks for enterprise insurance (such as employment liability insurance, etc.), they need to obtain enterprise information, employee information, and the relationship between enterprises and employees to make judgments. Enterprise insurance risk assessment systems usually rely on rule engines and traditional statistical analysis methods, and only judge the risks of enterprises and employees through historical insurance policy data and fixed rules, and are unable to mine and process the complex relationship between enterprises and employees. Based on this, existing methods introduce machine learning algorithms to identify risks for enterprise insurance, which can capture the relationship between enterprises and employees more deeply. However, due to the limitations of machine learning algorithms in extracting features of enterprise information and employee information, the relationship between enterprises and employees is not fully mined, which affects the accuracy of enterprise insurance risk identification results. Therefore, how to improve the accuracy of enterprise insurance risk identification results has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] The main purpose of the embodiments of the present application is to propose a risk identification method and device, electronic device and storage medium for enterprise insurance, aiming to improve the accuracy of risk identification results for enterprise insurance.

[0004] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present application proposes a risk identification method for enterprise insurance, the method comprising:

[0005] Acquire multiple historical enterprise insurance policy information, multiple claims information, and multiple enterprise-related information corresponding to the target enterprise insurance, wherein each of the enterprise-related information includes enterprise attribute information of the insured enterprise, attribute information of historically insured employees, employment relationship information, and historical insurance claims records;

[0006] Associating the historical enterprise insurance policy information, the claim settlement information and the enterprise information to obtain a plurality of reference information, wherein the insured enterprises corresponding to the respective reference information are different from each other;

[0007] Input all the reference information into a pre-trained machine learning model, and output the risk assessment results of each reference information;

[0008] Constructing a target relationship graph based on all of the enterprise attribute information, the employee attribute information, the risk assessment results, and the employment relationship information, wherein the nodes in the target relationship graph include enterprise nodes and employee nodes associated with the target enterprise insurance, and there is an employment relationship between the interconnected nodes in the target relationship graph;

[0009] Abnormal enterprise information and abnormal employee information are determined based on preset risk identification rules and the target relationship map, and the abnormal enterprise information and the abnormal employee information are determined as target risk identification results.

[0010] In some embodiments, the risk assessment result includes a first assessment result and a second assessment result, the first assessment result is used to indicate the active risk assessment result of the insured enterprise, and the second assessment result is used to indicate the active risk assessment result of the insured employee. A target relationship graph is constructed based on all of the enterprise attribute information, the employee attribute information, the risk assessment result, and the employment relationship information, including:

[0011] An initial relationship graph is constructed based on all of the enterprise attribute information, the employee attribute information, the risk assessment result, and the employment relationship information, wherein the initial relationship graph includes the enterprise node and the employee node, the enterprise node corresponds to a plurality of first node attributes, the first node attributes are obtained based on the corresponding enterprise attribute information and the first assessment result, and the attribute types of the first node attributes are different from each other, and the employee node corresponds to a plurality of second node attributes, the second node attributes are obtained based on the corresponding employee attribute information and the second assessment result, and the attribute types of the second node attributes are different from each other;

[0012] Determining first attribute weights of different first node attributes;

[0013] Determining second attribute weights of different second node attributes;

[0014] Based on a preset community discovery algorithm, each of the first attribute weights and each of the second attribute weights, the initial relationship graph is divided into communities and aggregated to obtain the target relationship graph.

[0015] In some embodiments, the initial relationship graph is divided into communities and aggregated based on a preset community discovery algorithm, each enterprise node weight, and each employee node weight to obtain the target relationship graph, including:

[0016] Performing community division on the initial relationship graph to obtain a first intermediate relationship graph, wherein the first intermediate relationship graph includes a plurality of initial communities, wherein each of the initial communities includes one of the enterprise nodes and the employee nodes having an employment relationship with the enterprise node;

[0017] Based on each of the first attribute weights and each of the second attribute weights, the enterprise nodes or the employee nodes in each of the initial communities are iteratively classified into each of the candidate communities to obtain a second intermediate relationship graph, wherein the second intermediate relationship graph includes a plurality of intermediate communities, wherein each of the initial communities corresponds to a plurality of the candidate communities, and the candidate communities of each of the initial communities are communities other than the initial community in the intermediate relationship graph;

[0018] Calculate the reference modularity of each intermediate community after each iteration operation;

[0019] Based on the reference modularity, calculating the modularity gain value of each intermediate community after each iterative operation;

[0020] When all the modularity gain values ​​are less than or equal to a preset gain threshold, the iterative operation is stopped to obtain the target relationship graph.

[0021] In some embodiments, the reference modularity is calculated according to the following formula:

[0022]

[0023] Among them, A ij is the connection weight between node i and node j, k i is the degree of node i, k j is the degree of node j, m is the total number of edges in the initial relationship graph, δ(c i ,c j ) is the Kronecker delta function. When nodes i and j belong to the same community, δ(c i ,c j ) takes the value of 1, when node i and node j do not belong to the same community, δ(c i ,c j ) takes a value of 0.

[0024] In some embodiments, the modularity gain value is calculated according to the following formula:

[0025] ΔQ=[∑in+k i ,in-(∑tot+k i ) 2 ]-[∑in-(∑tot) 2 ]-k i2 ;

[0026] Among them, ∑in is the weight sum of the edges of the corresponding intermediate community, ∑tot is the sum of the degrees of all nodes in the corresponding intermediate community, and k i is the degree of node i that is transferred from the corresponding intermediate community to the candidate community, k i ,in is the node weight of node i.

[0027] In some embodiments, determining abnormal enterprise information and abnormal employee information based on preset risk identification rules and the target relationship map includes:

[0028] Acquire multiple reference data sets from a reference database, each of the reference data sets including reference enterprise information associated with the insurance of the target enterprise and reference employee information associated with the enterprise corresponding to the reference enterprise information, wherein the reference enterprise information is used to indicate attribute information of an insured enterprise with insurance risk, and each of the reference employee information is used to indicate attribute information of an insured employee with insurance risk;

[0029] Obtaining first node attributes of enterprise nodes, second node attributes of employee nodes, and the number of employee nodes in each first target community in the target relationship graph;

[0030] Performing corresponding similarity matching between the first node attribute and the second node attribute in each first target community and the reference enterprise information and the reference employee information in each reference data set to obtain a plurality of similarity values;

[0031] Determine a second target community from the plurality of first target communities based on the similarity value, the number of employee nodes, a similarity threshold, and a range of the number of nodes, wherein the similarity value corresponding to the second target community is greater than or equal to the similarity threshold, and the number of employee nodes in the second target community exceeds the range of the number of nodes;

[0032] generating the abnormal enterprise information based on each first node attribute corresponding to the enterprise node in the second target community;

[0033] The abnormal employee information is generated based on each second node attribute corresponding to the employee node in the second target community.

[0034] In some embodiments, before dividing the initial relationship graph into communities based on a preset community discovery algorithm, each of the first attribute weights, and each of the second attribute weights, the method further includes:

[0035] When it is detected that the historical enterprise policy information, the claim settlement information or the enterprise-related information corresponding to the target enterprise insurance has changed, obtaining the changed new historical enterprise policy information, new claim settlement information or new enterprise-related information;

[0036] The node attribute value of the corresponding node in the initial relationship graph is updated based on the new historical enterprise policy information, the new claim settlement information or the new enterprise association information.

[0037] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application provides a risk identification device, the device comprising:

[0038] A data acquisition module, used to acquire a plurality of historical enterprise policy information, a plurality of claims information and a plurality of enterprise-related information corresponding to the target enterprise insurance, wherein each of the enterprise-related information includes enterprise attribute information of the insured enterprise, attribute information of historically insured employees, employment relationship information and historical insurance claims records;

[0039] A data processing module, used for associating the historical enterprise policy information, the claim settlement information and the enterprise information to obtain a plurality of reference information, wherein the insured enterprises corresponding to the respective reference information are different from each other;

[0040] A risk identification result acquisition module is used to input all the reference information into a pre-trained machine learning model and output the risk assessment results of each reference information;

[0041] A target relationship graph construction module, used to construct a target relationship graph based on all the enterprise attribute information, the employee attribute information, the risk assessment results and the employment relationship information, wherein the nodes in the target relationship graph include enterprise nodes and employee nodes associated with the target enterprise insurance, and there is an employment relationship between the interconnected nodes in the target relationship graph;

[0042] The risk identification result acquisition module is used to determine abnormal enterprise information and abnormal employee information based on preset risk identification rules and the target relationship map, and determine the abnormal enterprise information and the abnormal employee information as target risk identification results.

[0043] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, the memory stores a computer program, and the processor implements the risk identification method for enterprise insurance described in the first aspect when executing the computer program.

[0044] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the risk identification method for enterprise insurance described in the first aspect above.

[0045] The risk identification method and device, electronic device and storage medium of enterprise insurance proposed in the present application obtain multiple historical enterprise policy information, multiple claims information and multiple enterprise association information corresponding to the target enterprise insurance, wherein each of the enterprise association information includes enterprise attribute information of the insured enterprise, historical insured employee attribute information, employment relationship information and historical insurance claim records; associate the historical enterprise policy information, the claims information and the enterprise information to obtain multiple reference information, wherein the insured enterprises corresponding to each of the reference information are different; input all of the reference information into a pre-trained machine learning model, and output the risk assessment results of each of the reference information; construct a target relationship graph based on all of the enterprise attribute information, the employee attribute information, the risk assessment results and the employment relationship information, wherein the nodes in the target relationship graph include enterprise nodes and employee nodes associated with the target enterprise insurance, and there is an employment relationship between the interconnected nodes in the target relationship graph; determine abnormal enterprise information and abnormal employee information based on preset risk identification rules and the target relationship graph, and determine the abnormal enterprise information and the abnormal employee information as the target risk identification results. This application constructs a relationship map based on the data evaluation results obtained by the machine learning model, and combines the preset risk identification rules to deeply mine the relationships between nodes in the relationship map, thereby identifying abnormal corporate information and abnormal employee information, which can effectively improve the accuracy of risk identification results of corporate insurance. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a flow chart of a risk identification method for enterprise insurance provided by an embodiment of the present application;

[0047] Figure 2 yes Figure 1 Flow chart of step S40 in FIG.

[0048] Figure 3 yes Figure 2 Flow chart of step S44 in;

[0049] Figure 4 yes Figure 1 Flow chart of step S50 in FIG.

[0050] Figure 5 Yes Execution Figure 2 The flowchart before step S24 in FIG.

[0051] Figure 6 is a schematic diagram of a module of a risk identification device provided in an embodiment of the present application;

[0052] Figure 7 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0054] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0056] First, some nouns involved in this application are analyzed:

[0057] Artificial intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. AI is a branch of computer science that attempts to understand the essence of intelligence and produce a new type of intelligent machine that can respond in a similar way to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing and expert systems. AI can simulate the information process of human consciousness and thinking. AI is also a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0058] Machine Learning: An important branch of the field of artificial intelligence that focuses on using algorithms and statistical models to enable computer systems to automatically learn and improve from data without explicit programming. The core idea is to use data to train computer algorithms so that they can automatically learn from data and improve their performance without explicit programming. This process relies on the automatic identification and interpretation of data features in order to understand and solve complex practical problems. Machine learning algorithms are able to process large amounts of data, identify patterns, and make decisions or predictions based on these patterns by building and analyzing models. The basic principle can be summarized as: using algorithms to extract features from large amounts of data, building models, and then applying these models to predict or classify new data.

[0059] Information Extraction: A text processing technology that extracts specified types of entity, relationship, event and other factual information from natural language text and forms structured data output. Information extraction is a technology that extracts specific information from text data. Text data is composed of some specific units, such as sentences, paragraphs, and chapters. Text information is composed of some small specific units, such as characters, words, phrases, sentences, paragraphs, or a combination of these specific units. Extracting noun phrases, names, place names, etc. from text data are all text information extraction. Of course, the information extracted by text information extraction technology can be various types of information.

[0060] Relationship graph is a new type of knowledge representation and data management method. It displays entities and their relationships in the form of graphs, so as to better capture the connections and patterns between data. The construction of relationship graph requires the collection of a large amount of data, and the extraction of entities, relationships and attributes through technical means such as data mining and information extraction. Then, using graph databases or graph processing tools, these elements are connected in the form of graphs to form a complete relationship network. Elements such as entities, events, and behaviors are connected in the form of nodes and edges to form a complex relationship network. This network can clearly show the potential associations, patterns and laws between entities.

[0061] When enterprise insurance risk assessment systems identify risks for enterprise insurance (such as employment liability insurance, etc.), they need to obtain enterprise information, employee information, and the relationship between enterprises and employees to make judgments. Enterprise insurance risk assessment systems usually rely on rule engines and traditional statistical analysis methods, and only judge the risks of enterprises and employees through historical insurance policy data and fixed rules, and are unable to mine and process the complex relationship between enterprises and employees. Based on this, existing methods introduce machine learning algorithms to identify risks for enterprise insurance, which can capture the relationship between enterprises and employees more deeply. However, due to the limitations of machine learning algorithms in extracting features of enterprise information and employee information, the relationship between enterprises and employees is not fully mined, which affects the accuracy of enterprise insurance risk identification results. Therefore, how to improve the accuracy of enterprise insurance risk identification results has become a technical problem that needs to be solved urgently.

[0062] To solve the above-mentioned problems, the embodiments of the present application provide a risk identification method and device for enterprise insurance, an electronic device and a storage medium, which obtain multiple historical enterprise policy information, multiple claims information and multiple enterprise-related information corresponding to the target enterprise insurance, wherein each of the enterprise-related information includes enterprise attribute information of the insured enterprise, historical insured employee attribute information, employment relationship information and historical insurance claims records; associate the historical enterprise policy information, the claims information and the enterprise information to obtain multiple reference information, wherein the insured enterprises corresponding to each of the reference information are different; input all of the reference information into a pre-trained machine learning model, and output the risk assessment results of each of the reference information; construct a target relationship graph based on all of the enterprise attribute information, the employee attribute information, the risk assessment results and the employment relationship information, wherein the nodes in the target relationship graph include enterprise nodes and employee nodes associated with the target enterprise insurance, and there is an employment relationship between the interconnected nodes in the target relationship graph; determine abnormal enterprise information and abnormal employee information based on a preset risk identification rule and the target relationship graph, and determine the abnormal enterprise information and the abnormal employee information as the target risk identification result. This application constructs a relationship map based on the data evaluation results obtained by the machine learning model, and identifies abnormal corporate information and abnormal employee information from the relationship map in combination with preset risk identification rules. It can effectively improve the accuracy of risk identification results of corporate insurance, and has a high degree of automation, reducing manual intervention and improving identification efficiency.

[0063] The risk identification method and device, electronic device and storage medium for enterprise insurance provided in the embodiments of the present application are specifically illustrated through the following embodiments. First, the recommended method in the embodiments of the present application is described.

[0064] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0065] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0066] The risk identification method for enterprise insurance provided in the embodiment of the present application relates to the field of artificial intelligence technology. The risk identification method for enterprise insurance provided in the embodiment of the present application can be applied to a terminal, can be applied to a server side, or can be software running in a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or can be configured as a server cluster or a distributed system composed of multiple physical servers, or can be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the risk identification method for enterprise insurance, etc., but is not limited to the above forms.

[0067] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0068] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0069] Figure 1 is an optional flow chart of the risk identification method for enterprise insurance provided in the embodiment of the present application. Figure 1 The method may include but is not limited to steps S10 to S50.

[0070] Step S10, obtaining multiple historical enterprise policy information, multiple claims information and multiple enterprise association information corresponding to the target enterprise insurance, wherein each enterprise association information includes enterprise attribute information of the insured enterprise, historical insured employee attribute information, employment relationship information and historical insurance claims records.

[0071] Specifically, the target enterprise insurance of this embodiment is insurance for the enterprise and its employees, such as employment liability insurance, work-related injury insurance, etc.

[0072] Specifically, the risk identification for target enterprise insurance in this embodiment is to identify the information corresponding to the enterprise and its employees that have insured the target enterprise insurance, and whether there is any abnormality in the employment relationship between the enterprise and the employees, thereby protecting the rights and interests of the insurance company.

[0073] Specifically, the historical enterprise insurance policy information of this embodiment is the information agreed upon between the insurer and the insured, including the insured information (ie, the insured enterprise), the insured information (ie, the insured employee), the insurance subject and insurance liability information, and so on.

[0074] Specifically, claims information refers to the claims status of the target enterprise insurance for each historical enterprise policy information, including claims or no claims, which can provide an effective data basis for the subsequent identification of risks of insured enterprises and insured employees.

[0075] Specifically, the enterprise association information of this embodiment is used to indicate all enterprise attribute information of the current insured target enterprise insurance, as well as attribute information of historical insured employees who have an employment relationship with the insured enterprise, employment relationship information and historical insurance claim records.

[0076] Specifically, the enterprise attribute information of this embodiment includes industry type, corporate legal person name, corporate registration address, industry scale, organizational structure, etc.; employee attribute information includes employee name, employee ID number, employee length of service, employee position, employee responsibilities, employee job type, etc.

[0077] It can be understood that the historical enterprise policy information, claims information and enterprise-related information of this embodiment are scattered in different systems, and need to be obtained from different data sources respectively and fused. Obtaining the above information can provide an effective data basis for subsequent preliminary risk assessment and the construction of relationship maps.

[0078] Step S20, associating historical enterprise insurance policy information, claim settlement information and enterprise information to obtain a plurality of reference information, wherein the insured enterprises corresponding to the respective reference information are different from each other.

[0079] It should be noted that before associating historical corporate policy information, claims information and corporate information, the risk identification method of this embodiment also includes data cleaning of historical corporate policy information, claims information and corporate information, which can eliminate abnormal data such as corporate credit code 0, thereby ensuring the accuracy of subsequent risk identification results.

[0080] It is understandable that the historical enterprise insurance policy information, claims information and enterprise information obtained from multiple different data sources involve different enterprises and employee information of the target enterprise insurance. The target enterprise insurance is between the insurance company and the employees with an employment relationship. Therefore, it is necessary to associate and integrate these data to obtain multiple reference information. Each reference information corresponds to a different insured enterprise. Each reference information is associated with the information of an enterprise and its employees and the target enterprise insurance, thereby providing an effective data basis for subsequent risk identification.

[0081] Step S30, input all reference information into a pre-trained machine learning model, and output the risk assessment results of each reference information.

[0082] Specifically, this embodiment does not involve improvements to the model. The trained machine learning model can be a LightGBM model. The training data is the associated reference information with labels. The labels in this embodiment are used to indicate the claim results (whether there is a claim) corresponding to the insured objects of the target enterprise insurance. If the claim result is a rejection, there is a risk, and if the claim result is a claim, there is no risk, thereby providing an effective data basis for improving the accuracy of the subsequently trained machine learning model and ensuring the accuracy of the risk assessment results.

[0083] Specifically, this embodiment does not limit the specific number of training sets and test sets for training the machine learning model. It can be 18 million reference information for a preset year, of which 16 million reference information are training sets and 200 reference information are test sets. After an intermediate model is obtained by training the training set, the accuracy and recall rate of the intermediate model are tested using the test set and the corresponding true labels. When the accuracy meets the preset conditions, for example, the auc value is greater than or equal to 90% and the test accuracy is 53%, the model iteration training operation is stopped to obtain a trained machine learning model.

[0084] Step S40, constructing a target relationship graph based on all enterprise attribute information, employee attribute information, risk assessment results and employment relationship information, wherein the nodes in the target relationship graph include enterprise nodes and employee nodes associated with the target enterprise insurance, and there is an employment relationship between the interconnected nodes in the target relationship graph.

[0085] It can be understood that after obtaining the model evaluation results in step S30 of this embodiment, a relationship map is constructed based on the risk assessment results and enterprise attribute information, employee attribute information and employment relationship information, which increases the types of information for constructing the target relationship map, can enhance the complexity and depth of the map, can reveal more levels of associations and patterns, can improve the comprehensiveness and accuracy of the information, and provide an effective data basis for subsequent accurate risk identification results.

[0086] Step S50, determining abnormal enterprise information and abnormal employee information based on preset risk identification rules and the target relationship map, and determining the abnormal enterprise information and abnormal employee information as target risk identification results.

[0087] It should be noted that the risk identification rules of this embodiment can be to extract the characteristic information of each node in the target relationship graph (i.e., enterprise characteristic information and employee characteristic information with employment relationship), compare the characteristic information with the abnormal characteristic information in a preset database, and determine the abnormal enterprise information and abnormal employee information based on the comparison results; it can also be to perform cluster analysis on the target relationship graph to determine the abnormal enterprise information and abnormal employee information, or to use a machine learning model to analyze and determine the data of the target relationship graph.

[0088] See also Figure 2 In some embodiments, step S40 may include but is not limited to steps S41 to S44:

[0089] Step S41, constructing an initial relationship graph based on all enterprise attribute information, employee attribute information, risk assessment results and employment relationship information, wherein the initial relationship graph includes enterprise nodes and employee nodes, the enterprise nodes correspond to a plurality of first node attributes, the first node attributes are obtained based on the corresponding enterprise attribute information and the first assessment result, and the attribute types of the first node attributes are different from each other, and the employee nodes correspond to a plurality of second node attributes, the second node attributes are obtained based on the corresponding employee attribute information and the second assessment result, and the attribute types of the second node attributes are different from each other;

[0090] Step S42, determining first attribute weights of different first node attributes;

[0091] Step S43, determining second attribute weights of different second node attributes;

[0092] Step S44, based on the preset community discovery algorithm, each first attribute weight and each second attribute weight, the initial relationship graph is divided into communities and aggregated to obtain a target relationship graph.

[0093] Specifically, the embodiments of the present application do not limit the specific community discovery algorithm, which may be a label propagation algorithm, a hierarchical clustering algorithm, or a Louvain algorithm, and those skilled in the art may select one according to actual needs.

[0094] Specifically, after constructing the initial relationship graph, the node attributes of the enterprise nodes and employee nodes in the initial relationship graph of this embodiment have corresponding risk assessment results, which are used to indicate whether the enterprise node or employee node of the insured target enterprise insurance is rejected for compensation or whether the loss is cleared. If the risk assessment result is rejected for compensation or cleared for loss, it means that the corresponding enterprise node or employee node is at risk.

[0095] Specifically, the first node attributes include industry type, industry scale, organizational structure, corporate legal person, corporate credit code, corporate registration address, corporate telephone number, risk assessment results, etc.; the second node attributes include employee name, employee position, employee responsibilities, employee job type, employee length of service, employee gender, employee ID number, risk assessment results, etc.

[0096] Specifically, by dividing the initial relationship graph through the community discovery algorithm, at least one community with a close connection relationship can be found, which is helpful to understand the structure and function of the entire network. Within the community, the interaction between nodes is more frequent, and they may share similar attributes, interests or functions. By identifying these communities, we can have a deeper understanding of the local characteristics and global characteristics of the network. In addition, before dividing the initial relationship graph, this embodiment pre-assigns weights to different types of node attributes in different types of nodes, so that the community division result of the target relationship graph after community division meets the preset requirements. For example, this embodiment sets the industry type (whether it is a high-risk high-risk industry type), corporate legal person name, legal person address, and risk assessment result corresponding to the first node attribute corresponding to the enterprise node to have a higher attribute weight than the other first node attributes (such as corporate phone, structure organization); and the employee ID number, employee length of service and risk assessment result corresponding to the second node attribute corresponding to the employee node have a higher attribute weight than the other second node attributes (such as employee responsibilities, employee job type), which can make the risky enterprise nodes and their employee nodes allocated in the same community, providing effective support for subsequent risk identification.

[0097] See also Figure 5 In some embodiments, before executing step S24, the risk identification method for enterprise insurance of this embodiment further includes but is not limited to steps S241 to S242:

[0098] Step S241, when it is detected that the historical enterprise policy information, claim settlement information or enterprise-related information corresponding to the target enterprise insurance has changed, new historical enterprise policy information, new claim settlement information or new enterprise-related information after the change is obtained;

[0099] Step S242, based on new historical enterprise policy information, new claims information or new enterprise association information, update the node attribute value of the corresponding node in the initial relationship graph.

[0100] It is understandable that existing risk identification solutions for corporate insurance mainly rely on static historical data for analysis, and are unable to update and capture the latest employment relationships and risks in real time. Based on this, before dividing the initial relationship graph into communities, this embodiment detects whether the historical corporate policy information, claims information, or corporate association information corresponding to the target corporate insurance has changed (for example, the employee corresponding to employee node A resigns from the company corresponding to company node B and joins the company corresponding to company node C). When a change is detected, the new historical corporate policy information, new claims information, or new corporate association information after the change is obtained. Based on the new historical corporate policy information, new claims information, or new corporate association information, the node attribute value of the corresponding node is partially updated, thereby being able to update and capture the employment relationship and risks of the corporate node or employee node in real time, thereby ensuring the accuracy of the risk identification results.

[0101] See also Figure 3 In some embodiments, step S44 may include but is not limited to steps S441 to S445:

[0102] Step S441, dividing the initial relationship graph into communities to obtain a first intermediate relationship graph, the first intermediate relationship graph including a plurality of initial communities, wherein each initial community includes an enterprise node and an employee node having an employment relationship with the enterprise node;

[0103] Step S442: based on each first attribute weight and each second attribute weight, the enterprise nodes or employee nodes in each initial community are iteratively classified into each candidate community to obtain a second intermediate relationship graph, wherein the second intermediate relationship graph includes multiple intermediate communities, wherein each initial community corresponds to multiple candidate communities, and the candidate communities of each initial community are communities other than the initial community in the intermediate relationship graph;

[0104] Step S443, calculating the reference modularity of each intermediate community after each iteration operation;

[0105] Step S444, based on the reference modularity, calculating the modularity gain value of each intermediate community after each iteration operation;

[0106] Step S445, when all modularity gain values ​​are less than or equal to a preset gain threshold, the iteration operation is stopped to obtain a target relationship graph.

[0107] Specifically, the community discovery algorithm of this embodiment is the Louvain algorithm. Step S44, i.e., performing community division and community aggregation on the initial relationship graph based on the preset community discovery algorithm, each first attribute weight and each second attribute weight, to obtain the target relationship graph comprises: performing community division on the initial relationship graph to obtain a first intermediate relationship graph, the first intermediate relationship graph comprising multiple initial communities, wherein each initial community comprises an enterprise node and an employee node having an employment relationship with the enterprise node; based on each first attribute weight and each second attribute weight, the enterprise nodes or employee nodes in each initial community are iteratively classified into each candidate community to obtain a second intermediate relationship graph, the second intermediate relationship graph comprising multiple intermediate communities, wherein each initial community corresponds to multiple candidate communities, and the candidate communities of each initial community are communities other than the initial community in the intermediate relationship graph; after calculating each iteration operation, the reference model of each intermediate community is calculated. Blockiness; Based on the reference modularity, the modularity gain value of each intermediate community after each iterative operation is calculated. It can be understood that the modularity gain value refers to the change in modularity when we move a node or attribute from one intermediate community to another intermediate community. It indicates whether this movement helps to improve the quality of community division of the network; if moving a node or attribute to a new intermediate community results in a positive modularity gain, it means that this movement increases the connection density within the community (or reduces the connection across communities), thereby improving the community structure of the network; if the movement results in a negative modularity gain, it means that this movement reduces the connection density within the community (or increases the connection across communities), weakening the community structure of the network. After multiple iterative operations, when it is detected that all modularity gain values ​​are less than or equal to the preset gain threshold, for example, the preset gain threshold is 0, that is, when it is detected that the modularity gain value no longer increases significantly, the iterative operation is stopped to obtain the target relationship map.

[0108] Specifically, the reference modularity of this embodiment is calculated according to the following formula:

[0109]

[0110] Among them, A ij is the connection weight between node i and node j, k i is the degree of node i, k j is the degree of node j, m is the total number of edges in the initial relationship graph, δ(c i ,c j ) is the Kronecker delta function. When nodes i and j belong to the same community, δ(c i ,c j ) takes the value of 1, when node i and node j do not belong to the same community, δ(c i ,c j) takes a value of 0.

[0111] Specifically, the modularity gain value of this embodiment is calculated according to the following formula:

[0112] ΔQ=[∑in+k i ,in-(∑tot+k i ) 2 ]-[∑in-(∑tot) 2 ]-k i 2 ;

[0113] Among them, ∑in is the weight sum of the edges of the corresponding intermediate community, ∑tot is the sum of the degrees of all nodes in the corresponding intermediate community, and k i is the degree of node i that is transferred from the corresponding intermediate community to the candidate community, k i ,in is the node weight of node i. The node weight is obtained by summing the attribute weights corresponding to node i. The modularity gain value can also be calculated by the difference between the new and old reference modularities (i.e., the difference between the reference modularities before and after the node or attribute moves to the new intermediate community).

[0114] See also Figure 4 In some embodiments, step S50 may include but is not limited to steps S51 to S56:

[0115] Step S51, obtaining multiple reference data sets from a reference database, each reference data set including reference enterprise information associated with the insurance of the target enterprise and reference employee information associated with the enterprise corresponding to the reference enterprise information, wherein the reference enterprise information is used to indicate attribute information of the insured enterprise with insurance risk, and each reference employee information is used to indicate attribute information of the insured employee with insurance risk;

[0116] Step S52, obtaining the first node attribute of the enterprise node, the second node attribute of the employee node and the number of employee nodes in each first target community in the target relationship graph;

[0117] Step S53, performing corresponding similarity matching between the first node attribute and the second node attribute in each first target community and the reference enterprise information and the reference employee information in each reference data set to obtain a plurality of similarity values;

[0118] Step S54, determining a second target community from the plurality of first target communities based on the similarity value, the number of employee nodes, the similarity threshold and the range of the number of nodes, wherein the similarity value corresponding to the second target community is greater than or equal to the similarity threshold, and the number of employee nodes in the second target community exceeds the range of the number of nodes;

[0119] Step S55, generating abnormal enterprise information based on each first node attribute corresponding to the enterprise node in the second target community;

[0120] Step S56: generating abnormal employee information based on each second node attribute corresponding to the employee node in the second target community.

[0121] Specifically, the reference database of this embodiment stores real enterprise data and employee data that are at risk within a preset time period. The reference database stores data of the same type as the attribute values ​​of each node in the target relationship graph, which can provide effective data support for identifying whether the insured enterprises and insured employees corresponding to the current insured target enterprise insurance are at risk.

[0122] Specifically, step S50 of this embodiment includes the following steps: obtaining multiple reference data sets from a reference database, each reference data set including reference enterprise information associated with the insurance of the target enterprise and reference employee information associated with the enterprise corresponding to the reference enterprise information, wherein the reference enterprise information is used to indicate attribute information of the insured enterprise with insurance risk, and each reference employee information is used to indicate attribute information of the insured employee with insurance risk; obtaining the first node attribute of the enterprise node, the second node attribute of the employee node and the number of employee nodes in each first target community in the target relationship graph; comparing the first node attribute and the second node attribute in each first target community with the reference enterprise information and the reference employee information in each reference data set; The method further comprises the steps of: performing corresponding similarity matching on the employee information of the enterprise nodes and obtaining multiple similarity values; determining a second target community from multiple first target communities based on the similarity value, the number of employee nodes, the similarity threshold and the range of the number of nodes, wherein the similarity value corresponding to the second target community is greater than or equal to the similarity threshold, and the number of employee nodes in the second target community exceeds the range of the number of nodes; generating abnormal enterprise information based on each first node attribute corresponding to the enterprise node in the second target community; generating abnormal employee information based on each second node attribute corresponding to the employee node in the second target community. It can be understood that the abnormal enterprise information and abnormal employee information identified according to the present solution can help the business end to accurately evaluate and identify the risks of the target enterprise insurance.

[0123] Specifically, the similarity threshold and the range of the number of nodes in this embodiment can be adjusted by those skilled in the art according to actual conditions. The risk level of abnormal enterprise information and abnormal employee information can also be determined by setting different similarity thresholds and ranges of the number of nodes. For example, the similarity threshold is 60%, the range of the number of nodes is [5,100], the similarity value corresponding to the second target community 1 is greater than or equal to 60%, and the number of employee nodes is 120 (outside the range of the number of nodes), and the similarity value corresponding to the second target community 2 is greater than or equal to 60%, and the number of employee nodes is 56 (within the range of the number of nodes), then the risk level of the second target community 1 is higher than the risk level of the second target community 2.

[0124] In addition, after executing step S50, this embodiment also includes generating prompt information based on the target risk identification result (including that the number of abnormal enterprises that have insured the target enterprise within a preset period is x, accounting for x1% of the current total number of enterprises, the corresponding number of employees is y, and a specific list of abnormal enterprises and a list of abnormal employees) and sending it to the business end so that the business end can perform risk management.

[0125] See also Figure 6 The embodiment of the present application also provides a risk identification device, which can implement the above-mentioned enterprise insurance risk identification method, and the device includes:

[0126] A data acquisition module is used to acquire multiple historical enterprise policy information, multiple claims information and multiple enterprise-related information corresponding to the target enterprise insurance, wherein each enterprise-related information includes enterprise attribute information of the insured enterprise, attribute information of historically insured employees, employment relationship information and historical insurance claims records;

[0127] A data processing module is used to associate historical enterprise insurance policy information, claim settlement information and enterprise information to obtain multiple reference information, wherein each reference information corresponds to a different insured enterprise;

[0128] The risk assessment module is used to input all reference information into a pre-trained machine learning model and output the risk assessment results of each reference information;

[0129] A target relationship graph construction module is used to construct a target relationship graph based on all enterprise attribute information, employee attribute information, risk assessment results and employment relationship information, wherein the nodes in the target relationship graph include enterprise nodes and employee nodes associated with the target enterprise insurance, and there is an employment relationship between the interconnected nodes in the target relationship graph;

[0130] The risk identification result acquisition module is used to determine abnormal enterprise information and abnormal employee information based on preset risk identification rules and target relationship maps, and determine the abnormal enterprise information and abnormal employee information as target risk identification results.

[0131] The specific implementation of the risk identification device is basically the same as the specific implementation of the risk identification method for enterprise insurance mentioned above, and will not be repeated here.

[0132] The embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned enterprise insurance risk identification method when executing the computer program. The electronic device can be any smart terminal including a tablet computer, a car computer, etc.

[0133] See also Figure 7 , Figure 7 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0134] The processor 701 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0135] The memory 702 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 702 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 702, and the processor 701 calls and executes the enterprise insurance risk identification method of the embodiment of this application;

[0136] Input / output interface 703, used to implement information input and output;

[0137] Communication interface 704, used to realize communication interaction between the device and other devices, which can be realized through wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0138] A bus 705 that transmits information between the various components of the device (e.g., the processor 701, the memory 702, the input / output interface 703, and the communication interface 704);

[0139] The processor 701 , the memory 702 , the input / output interface 703 and the communication interface 704 are connected to each other in communication within the device via a bus 705 .

[0140] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the risk identification method for enterprise insurance.

[0141] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0142] The enterprise insurance risk identification method and device, electronic device and storage medium provided in the embodiments of the present application obtain multiple historical enterprise policy information, multiple claims information and multiple enterprise-related information corresponding to the target enterprise insurance, wherein each enterprise-related information includes enterprise attribute information of the insured enterprise, historical insured employee attribute information, employment relationship information and historical insurance claim records; associate the historical enterprise policy information, claims information and enterprise information to obtain multiple reference information, wherein the insured enterprises corresponding to each reference information are different; input all the reference information into a pre-trained machine learning model, and output the risk assessment result of each reference information; construct a target relationship graph based on all the enterprise attribute information, employee attribute information, risk assessment results and employment relationship information, wherein the nodes in the target relationship graph include enterprise nodes and employee nodes associated with the target enterprise insurance, and there is an employment relationship between the nodes connected to each other in the target relationship graph; determine abnormal enterprise information and abnormal employee information based on preset risk identification rules and the target relationship graph, and determine the abnormal enterprise information and abnormal employee information as the target risk identification result. This application constructs a relationship map based on the data evaluation results obtained by the machine learning model, and deeply mines the relationship between nodes from the relationship map in combination with the preset risk identification rules, thereby identifying abnormal corporate information and abnormal employee information, which can effectively improve the accuracy of risk identification results of corporate insurance.

[0143] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0144] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0145] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0146] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0147] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0148] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0149] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0150] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0151] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0152] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, 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, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0153] The preferred embodiments of the present application are described above with reference to the accompanying drawings, but the scope of the rights of the present application is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present application should be within the scope of the rights of the present application.

Claims

1. A risk identification method for enterprise insurance, characterized in that: The method comprises: Acquire multiple historical enterprise insurance policy information, multiple claims information, and multiple enterprise-related information corresponding to the target enterprise insurance, wherein each of the enterprise-related information includes enterprise attribute information of the insured enterprise, attribute information of historically insured employees, employment relationship information, and historical insurance claims records; Associating the historical enterprise insurance policy information, the claim settlement information and the enterprise information to obtain a plurality of reference information, wherein the insured enterprises corresponding to the respective reference information are different from each other; Input all the reference information into a pre-trained machine learning model, and output the risk assessment results of each reference information; Constructing a target relationship graph based on all of the enterprise attribute information, the employee attribute information, the risk assessment results, and the employment relationship information, wherein the nodes in the target relationship graph include enterprise nodes and employee nodes associated with the target enterprise insurance, and there is an employment relationship between the interconnected nodes in the target relationship graph; Abnormal enterprise information and abnormal employee information are determined based on preset risk identification rules and the target relationship map, and the abnormal enterprise information and the abnormal employee information are determined as target risk identification results.

2. The risk identification method according to claim 1, characterized in that: The risk assessment result includes a first assessment result and a second assessment result, wherein the first assessment result is used to indicate the active risk assessment result of the insured enterprise, and the second assessment result is used to indicate the active risk assessment result of the insured employee. A target relationship graph is constructed based on all of the enterprise attribute information, the employee attribute information, the risk assessment result, and the employment relationship information, including: An initial relationship graph is constructed based on all of the enterprise attribute information, the employee attribute information, the risk assessment result, and the employment relationship information, wherein the initial relationship graph includes the enterprise node and the employee node, the enterprise node corresponds to a plurality of first node attributes, the first node attributes are obtained based on the corresponding enterprise attribute information and the first assessment result, and the attribute types of the first node attributes are different from each other, and the employee node corresponds to a plurality of second node attributes, the second node attributes are obtained based on the corresponding employee attribute information and the second assessment result, and the attribute types of the second node attributes are different from each other; Determining first attribute weights of different first node attributes; Determining second attribute weights of different second node attributes; Based on a preset community discovery algorithm, each of the first attribute weights and each of the second attribute weights, the initial relationship graph is divided into communities and aggregated to obtain the target relationship graph.

3. The risk identification method according to claim 2, characterized in that: Based on a preset community discovery algorithm, each enterprise node weight and each employee node weight, the initial relationship graph is divided into communities and aggregated to obtain the target relationship graph, including: Performing community division on the initial relationship graph to obtain a first intermediate relationship graph, wherein the first intermediate relationship graph includes a plurality of initial communities, wherein each of the initial communities includes one of the enterprise nodes and the employee nodes having an employment relationship with the enterprise node; Based on each of the first attribute weights and each of the second attribute weights, the enterprise nodes or the employee nodes in each of the initial communities are iteratively classified into each of the candidate communities to obtain a second intermediate relationship graph, wherein the second intermediate relationship graph includes a plurality of intermediate communities, wherein each of the initial communities corresponds to a plurality of the candidate communities, and the candidate communities of each of the initial communities are communities other than the initial community in the intermediate relationship graph; Calculate the reference modularity of each intermediate community after each iteration operation; Based on the reference modularity, calculating the modularity gain value of each intermediate community after each iterative operation; When all the modularity gain values ​​are less than or equal to a preset gain threshold, the iteration operation is stopped to obtain the target relationship graph.

4. The risk identification method according to claim 3, characterized in that: The reference modularity is calculated according to the following formula: Among them, A ij is the connection weight between node i and node j, k i is the degree of node i, k j is the degree of node j, m is the total number of edges in the initial relationship graph, δ(c i ,c j ) is the Kronecker delta function. When nodes i and j belong to the same community, δ(c i ,c j ) takes the value of 1, when node i and node j do not belong to the same community, δ(c i ,c j ) takes a value of 0.

5. The risk identification method according to claim 3, characterized in that: The modularity gain value is calculated according to the following formula: ΔQ=[∑in+k i ,in-(∑tot+k i ) 2 ]-[∑in-(∑tot) 2 ]-k i 2 ; Among them, ∑in is the weight sum of the edges of the corresponding intermediate community, ∑tot is the sum of the degrees of all nodes in the corresponding intermediate community, and k i is the degree of node i that is transferred from the corresponding intermediate community to the candidate community, k i ,in is the node weight of node i.

6. The risk identification method according to claim 2, characterized in that: Determining abnormal enterprise information and abnormal employee information based on preset risk identification rules and the target relationship map includes: Acquire multiple reference data sets from a reference database, each of the reference data sets including reference enterprise information associated with the insurance of the target enterprise and reference employee information associated with the enterprise corresponding to the reference enterprise information, wherein the reference enterprise information is used to indicate attribute information of an insured enterprise with insurance risk, and each of the reference employee information is used to indicate attribute information of an insured employee with insurance risk; Obtaining first node attributes of enterprise nodes, second node attributes of employee nodes, and the number of employee nodes in each first target community in the target relationship graph; Performing corresponding similarity matching between the first node attribute and the second node attribute in each first target community and the reference enterprise information and the reference employee information in each reference data set to obtain a plurality of similarity values; Determine a second target community from the plurality of first target communities based on the similarity value, the number of employee nodes, a similarity threshold, and a range of the number of nodes, wherein the similarity value corresponding to the second target community is greater than or equal to the similarity threshold, and the number of employee nodes in the second target community exceeds the range of the number of nodes; generating the abnormal enterprise information based on each first node attribute corresponding to the enterprise node in the second target community; The abnormal employee information is generated based on each second node attribute corresponding to the employee node in the second target community.

7. The risk identification method according to claim 2, characterized in that: Before dividing the initial relationship graph into communities based on a preset community discovery algorithm, each of the first attribute weights, and each of the second attribute weights, the method further includes: When it is detected that the historical enterprise policy information, the claim settlement information or the enterprise-related information corresponding to the target enterprise insurance has changed, obtaining the changed new historical enterprise policy information, new claim settlement information or new enterprise-related information; The node attribute value of the corresponding node in the initial relationship graph is updated based on the new historical enterprise policy information, the new claim settlement information or the new enterprise association information.

8. A risk identification device, characterized in that: The device comprises: A data acquisition module, used to acquire a plurality of historical enterprise policy information, a plurality of claims information and a plurality of enterprise-related information corresponding to the target enterprise insurance, wherein each of the enterprise-related information includes enterprise attribute information of the insured enterprise, attribute information of historically insured employees, employment relationship information and historical insurance claims records; A data processing module, used for associating the historical enterprise policy information, the claim settlement information and the enterprise information to obtain a plurality of reference information, wherein the insured enterprises corresponding to the respective reference information are different from each other; A risk assessment module, used to input all the reference information into a pre-trained machine learning model and output a risk assessment result of each reference information; A target relationship graph construction module, used to construct a target relationship graph based on all the enterprise attribute information, the employee attribute information, the risk assessment results and the employment relationship information, wherein the nodes in the target relationship graph include enterprise nodes and employee nodes associated with the target enterprise insurance, and there is an employment relationship between the interconnected nodes in the target relationship graph; The risk identification result acquisition module is used to determine abnormal enterprise information and abnormal employee information based on preset risk identification rules and the target relationship map, and determine the abnormal enterprise information and the abnormal employee information as target risk identification results.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the enterprise insurance risk identification method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the risk identification method for enterprise insurance according to any one of claims 1 to 7 is implemented.

Images