Insurance policy processing method and device for group customers, storage medium and electronic equipment

Abstract

The invention discloses a group customer insurance policy processing method and device, a storage medium and electronic equipment. The invention belongs to the technical field of computers. The method comprises the following steps: acquiring industry labels and historical insurance data of group customers, and inputting the industry labels and the historical insurance data into a risk assessment module to obtain risk assessment results corresponding to the industry labels; calling a portrait construction module to obtain a group customer portrait corresponding to each industry label; when an insurance buying service is received, acquiring insurance buying main body information, and inputting the insurance buying main body information to an industry matching module to obtain a target industry label; and inputting the group user portrait of the target industry label and the historical insurance record of the insurance subject to an insurance service analysis module to obtain a processing result of the insurance service. According to the technical scheme, refined classification and accurate positioning analysis can be carried out for group customer insurance, insurance businesses of group customers are stably processed, and the processing efficiency and the processing accuracy of the insurance businesses are improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and particularly to a method, device, storage medium and electronic device for processing insurance policies of group customers. Background Art

[0002] With the stable development of the economy, the awareness of insurance among the general public is getting stronger and stronger. Among a large number of insurance demands, there is a special type of insurance demand, that is, group customers' insurance. After receiving an insurance request, currently, the claim query is still carried out with a single customer as the main body. For example, the historical insurance records of a certain individual in the group customer will be queried. And currently, the query object is only limited to the comprehensive loss ratio, lacking more refined analysis, so accurate risk analysis results cannot be obtained. Moreover, in some scenarios, the analysis tools are insufficient, lacking systematic analysis tools, and still in the mode of relying on reports, data extraction, and manual Excel calculations. Therefore, how to reasonably process the insurance demands of group customers is one of the difficult technical problems for those skilled in the art. Summary of the Invention

[0003] This application proposes a method, device, storage medium and electronic device for processing insurance policies of group customers, so as to solve the problems of inability to conduct reasonable risk analysis and insufficient analysis tools and means for group customers' insurance. The technical solution provided by this application can carry out refined classification and accurate positioning analysis for group customers' insurance, handle the insurance business of group customers steadily, improve the processing efficiency and accuracy of the insurance business, and enhance the customer experience.

[0004] An embodiment of this application provides a method for processing insurance policies of group customers. The method is executed by an insurance policy management system, and the insurance policy management system is deployed in an electronic device. The method includes:

[0005] Obtain the industry labels and historical insurance data of group customers, input the industry labels and the historical insurance data into a risk assessment module, and obtain the risk assessment results corresponding to each industry label;

[0006] Based on the risk assessment results corresponding to each industry label, call a portrait construction module to obtain group customer portraits corresponding to each industry label;

[0007] When receiving an insurance business, obtain the insurance subject information, input the insurance subject information into an industry matching module, and obtain the target industry label corresponding to the insurance subject information;

[0008] Input the group user portrait of the target industry label and the historical insurance records of the insured entity in the insured business analysis module to obtain the processing result of the insured business.

[0009] Further, obtain the industry labels and historical insurance data of group customers, and input the industry labels and the historical insurance data into the risk assessment module to obtain the risk assessment results corresponding to each industry label, including:

[0010] Obtain the customer information of group customers, and determine the industry labels of the group customers according to the customer information; wherein, the customer information includes one or more of industry classification, enterprise scale, operating conditions, and geographical location.

[0011] Obtain the historical insurance data of group customers, and determine the policy and claim information according to the historical insurance data; wherein, the historical insurance data includes one or more of insurance type, insurance amount, claim record, claim ratio, claim frequency, claim amount, and historical policy details.

[0012] According to the customer information and historical insurance data of the group customers, use the risk assessment module to obtain the risk assessment results corresponding to each industry label.

[0013] Further, the method further includes:

[0014] Establish a basic analysis model;

[0015] Use the customer information and historical insurance data of group customers as sample data to construct a training set and a test set respectively. After training the basic analysis model with the training set, if the risk assessment results of the test set using the trained analysis model meet the preset accuracy rate, it is determined that the analysis model training is completed.

[0016] Correspondingly, according to the customer information and historical insurance data of the group customers, use the risk assessment module to obtain the risk assessment results corresponding to each industry label, including:

[0017] Use the analysis model of the risk assessment module to analyze the customer information and historical insurance data of the group customers to obtain the risk assessment results corresponding to each industry label.

[0018] Further, using the customer information and historical insurance data of group customers as sample data includes:

[0019] Obtain the basic data of the customer information and historical insurance data of the group customers.

[0020] Clean the basic data to obtain the cleaned result data; wherein, the data cleaning includes one or more of removing duplicate records, handling missing values, filling specific values, and handling outliers;

[0021] Use the cleaned result data as sample data.

[0022] Further, the risk assessment module further includes a claim data integration model, and the training process of the claim data integration model includes:

[0023] Obtain the claim-related data in the historical insurance application data; wherein, the claim-related data includes one or more of claim amount, claim frequency, claim reason, claim time, policyholder information, and underwriting terms;

[0024] Establish the initial claim data integration model and use the claim-related data for training to obtain the claim data integration model; wherein, the claim data integration model is used to output at least one result of trend analysis, correlation analysis, risk scoring, and predictive analysis.

[0025] Further, after obtaining the claim-related data in the historical insurance application data, the method further includes:

[0026] Perform feature engineering processing on the claim-related data;

[0027] After using the claim-related data for training to obtain the claim data integration model, the method further includes:

[0028] Use cross-validation and / or A / B testing to evaluate the performance of the claim data integration model, and adjust the parameters of the claim data integration model or replace the claim data integration model according to the model performance evaluation results.

[0029] Further, the insurance application business analysis module includes a pricing model;

[0030] Wherein, the pricing model is used to determine the insurance application pricing result of the current insurance application business based on the group user portrait of the target industry label and the historical insurance application records of the insured entity in the insured entity information.

[0031] An embodiment of the present application further provides a policy processing device for group customers, the device is configured in a policy management system, and the policy management system is deployed in an electronic device; the device includes:

[0032] A risk assessment module, configured to obtain the industry label and historical insurance application data of a group customer, and determine the risk assessment result corresponding to each industry label according to the industry label and the historical insurance application data;

[0033] An image construction module, configured to generate a group customer image corresponding to each industry label based on the risk assessment results corresponding to the respective industry labels;

[0034] An industry matching module, configured to obtain the insured subject information when receiving an insurance application, match the insured subject information with the group customer images corresponding to the respective industry labels that have been obtained, and obtain a target industry label corresponding to the insured subject information;

[0035] An insurance application analysis module, configured to determine the processing result of the insurance application according to the group user image of the target industry label and the historical insurance records of the insured subject in the insured subject information.

[0036] An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned policy processing method for group customers is implemented.

[0037] An embodiment of the present application further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above-mentioned policy processing method for group customers is implemented.

[0038] The embodiment of the present application adopts the following technical solutions: obtaining the industry labels and historical insurance data of group customers, inputting the industry labels and the historical insurance data into a risk assessment module to obtain risk assessment results corresponding to the respective industry labels; based on the risk assessment results corresponding to the respective industry labels, calling an image construction module to obtain group customer images corresponding to the respective industry labels; obtaining the insured subject information when receiving an insurance application, inputting the insured subject information into an industry matching module to obtain a target industry label corresponding to the insured subject information; inputting the group user image of the target industry label and the historical insurance records of the insured subject in the insured subject information into an insurance application analysis module to obtain the processing result of the insurance application.

[0039] At least one of the above technical solutions adopted by the embodiment of the present application can achieve the following beneficial effects:

[0040] In the context of the rapid growth of group insurance for employee benefits, provide an accurate industry and customer underwriting analysis tool, improve the accuracy of product pricing, reduce the comprehensive claim ratio of policies, and effectively improve the underwriting quality of large customers;

[0041] Co-ordinate the management of short-term insurance combinations and optimize the pricing support for short-term insurance;

[0042] Rely on big data analysis to build a product claim assessment model;

[0043] Implement group management for key customers and optimize the pricing strategy for major customers;

[0044] Predict the company's expense creation using machine learning and artificial intelligence, and allocate resources reasonably;

[0045] Overall management of decentralized combined short-term insurance business, improve the level of intelligent analysis, effectively monitor the operating efficiency of the company's short-term insurance, and boost the scientific and intelligent operation decision-making of short-term insurance. Description of the Drawings

[0046] The drawings described herein are used to provide a further understanding of the present application, and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation of the present application. In the drawings:

[0047] Figure 1 It is a schematic flowchart of the policy processing method for group customers provided in Embodiment 1 of the present application;

[0048] Figure 2 It is a schematic structural diagram of the policy processing device for group customers provided in Embodiment 3 of the present application;

[0049] Figure 3 It is a schematic structural diagram of an electronic device provided in Embodiment 4 of the present application. Detailed Description of the Embodiments

[0050] To make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions of the present application will be clearly and completely described below in conjunction with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0051] The following will describe in detail the technical solutions provided by each embodiment of the present application in conjunction with the drawings.

[0052] Embodiment 1

[0053] Figure 1 It is a schematic flowchart of the policy processing method for group customers provided in Embodiment 1 of the present application. The method is executed by a policy management system, and the policy management system is deployed in an electronic device. As Figure 1 shown, the method includes:

[0054] S11, Obtain the industry labels and historical insurance data of group customers, input the industry labels and the historical insurance data into a risk assessment module, and obtain the risk assessment results corresponding to each industry label;

[0055] Among them, group customers can be customers who are a group entity consisting of multiple individuals, such as a large company that provides unified insurance for all its employees.

[0056] An industry label can be a mark used to identify the industry category to which the group's customers belong, such as "Internet technology industry", "financial services industry", "manufacturing industry", etc. It can help quickly classify customers by industry, making it easier for subsequent targeted analysis and other operations.

[0057] Historical insurance data may be the relevant data information retained during the group customers' past insurance policy purchase process, including the type of insurance purchased, the insured amount, the insurance claims, and other aspects. These data reflect the group customers' previous insurance experience and risk status.

[0058] Among them, the risk assessment module is a program module with specific functions. It can receive input data, here are industry labels and historical insurance data, analyze and process the input data according to the built-in algorithms and rules, and then output the corresponding risk assessment results. It can use the classification algorithm technology in machine learning, such as the decision tree algorithm, through learning and training a large amount of labeled historical similar data, to build a risk assessment model integrated into the module. The effect of this technology is that it can judge the risk situation corresponding to different industry labels based on the input data more accurately and efficiently, and reduce the subjectivity and errors of manual evaluation.

[0059] In this solution, data can be obtained by collecting industry labels and historical insurance data of group customers through corresponding data collection channels or system interfaces, so that they can be used in subsequent processes, such as obtaining industry label information from the enterprise registration database and extracting historical insurance data from the insurance business system database.

[0060] After the calculation, analysis and other processing processes within the risk assessment module, the risk assessment results corresponding to the corresponding industry labels are finally output, that is, the expected processing output content is obtained.

[0061] S12, based on the risk assessment results corresponding to the various industry labels, calling the portrait construction module to obtain the group customer portraits corresponding to the various industry labels;

[0062] The risk assessment results corresponding to each industry label are the evaluation contents on risk degree, risk type, etc. given for different industry labels, which are output by the risk assessment module in the previous step. For example, for the "manufacturing" industry label, the risk assessment results may show specific situations such as high risk in production safety and medium risk in property loss.

[0063] After receiving risk assessment results and other data, the portrait construction module can use data analysis, feature extraction and other technical means to integrate a portrait that can comprehensively reflect the characteristics of the group's customers in all aspects. For example, the knowledge graph technology can be used to associate and integrate data of different dimensions in a graph structure to construct a group customer portrait containing rich information. The technical effect is that it can intuitively and three-dimensionally display the comprehensive situation of the group's customers, making it easier to quickly understand the full picture of customer information of various industry labels in subsequent business applications.

[0064] The group customer portrait can be a visual or data-based presentation of the comprehensive characteristics of the group customer, covering many aspects such as the group customer's basic information, operating conditions, risk characteristics, and relevant characteristics of the insurance history.

[0065] S13, when receiving the insurance business, obtaining the insured subject information, inputting the insured subject information into the industry matching module, and obtaining the target industry label corresponding to the insured subject information;

[0066] Insurance business can be the commercial behavior of customers applying for insurance to insurance companies and a series of related operations and processes, such as a company purchasing property insurance for its fixed assets, group accident insurance for its employees and other specific insurance business matters.

[0067] The insured subject information is about the entity that submits the insurance application, which can be relevant information of the enterprise, organization, etc., such as the name, nature, business scope, scale and other basic information of the entity, as well as some past insurance-related records. For example, the insured subject information of an enterprise includes the enterprise's registered address, industry, previously insured types, etc. This information helps to accurately judge the situation of the insured subject.

[0068] The industry matching module is a functional module used to compare and match the insured subject information with the existing industry classification standards or industry label library. It can use fuzzy matching combined with weighted scoring algorithm technology to assign different weights to the key industry feature descriptions in the insured subject information, and then compare and score with the standard descriptions in the industry label library. The industry label with the highest matching degree is the result. Its technical effect is that it can determine the industry corresponding to the insured subject more flexibly and accurately, and adapt to insured subject information expressed in different ways.

[0069] The target industry label is obtained by matching the insured entity information. It represents the specific industry label of the industry to which the insured entity belongs. For example, after matching, it is determined that the target industry label of a certain enterprise is "transportation industry."

[0070] After processes such as matching operations within the industry matching module, this solution outputs the target industry label corresponding to the insured entity information, that is, it completes the operation from inputting the insured entity information to determining its industry label.

[0071] S14. Input the group user portrait of the target industry label and the historical insurance records of the insured entity in the insured entity information into the insurance business analysis module to obtain the processing result of the insurance business.

[0072] The historical insurance records of the insured entity in the insured entity information can be the detailed records specifically about its previous participation in insurance, such as how many times it has insured, the types of insurance for each insurance, the claim settlement situation, etc.

[0073] The insurance business analysis module is a functional module that can comprehensively evaluate and analyze the insurance business by using technologies such as data analysis and risk prediction, and then give corresponding processing results. It can adopt the technology of combining big data analysis with a risk prediction model to deeply mine and analyze the input portrait information, historical insurance records and other data. Its technical effect is to accurately judge the rationality and risk degree of the insurance business, and assist in business decision-making.

[0074] The processing result can be the final conclusion about the insurance business output by the insurance business analysis module after analyzing and processing various input information, such as whether to agree to underwrite, the underwriting conditions that need to be adjusted, risk warnings, etc. It is the final feedback given by the entire process for the insurance business.

[0075] After complex data analysis, operations, evaluation and other processes within the insurance business analysis module, the processing result of the insurance business is output, that is, the last step of comprehensively analyzing the insurance business and giving a conclusion in the entire process is completed.

[0076] This technical solution, through the coordinated operation of multiple functional modules, uses a variety of innovative technologies such as classification algorithms of machine learning, knowledge graphs, fuzzy matching combined with weight scoring, and big data analysis and risk prediction models. First, it accurately evaluates the risk situations corresponding to different industry labels, then constructs a comprehensive and intuitive group customer portrait, then accurately matches the industry to which the insured entity belongs, and finally deeply analyzes the insurance business and gives a reasonable processing result. Such a technical solution can effectively reduce the subjectivity and errors of manual operations, improve the efficiency and accuracy of data processing and business analysis, provide strong support for insurance business-related decisions in an all-round and three-dimensional manner, and at the same time can flexibly adapt to the diverse needs under different customers and business scenarios, greatly optimizing the entire process of insurance business from underwriting assessment to decision-making processing.

[0077] In one embodiment, optionally, obtain the industry tags and historical insurance data of group customers, and input the industry tags and the historical insurance data into a risk assessment module to obtain risk assessment results corresponding to each industry tag, including:

[0078] Obtain the customer information of group customers, and determine the industry tags of the group customers according to the customer information; wherein, the customer information includes one or more of industry classification, enterprise scale, operating conditions, and geographical location;

[0079] Obtain the historical insurance data of group customers, and determine policy and claim information according to the historical insurance data; wherein, the historical insurance data includes one or more of insurance type, insurance amount, claim record, claim ratio, claim frequency, claim amount, and historical policy details;

[0080] According to the customer information and historical insurance data of the group customers, use a risk assessment module to obtain risk assessment results corresponding to each industry tag.

[0081] Among them, the customer information may include: industry classification, such as belonging to specific industry fields such as manufacturing, service industries, etc., enterprise scale, like the division of different scale levels such as small, medium, and large enterprises, which can be measured by indicators such as assets and the number of employees, operating conditions, such as the state reflecting the operation of the enterprise including profit situation, business growth rate, market share, etc., and geographical location, such as the region where the enterprise headquarters is located, the regional scope of the main business carried out, etc. These information combined can help accurately locate the basic characteristics of group customers and the industry category they belong to, etc.

[0082] The policy and claim information can be the key content extracted from the historical insurance data. The insurance type refers to the specific insurance types that group customers have insured in the past, such as property insurance, life insurance, etc.; the insurance amount is the insured amount set for each insurance; the claim record shows whether there has been an insurance claim and the specific claim-related process; the claim ratio is the ratio of the claim amount to the insurance amount, reflecting the relative degree of insurance claims; the claim frequency reflects the number of insurance claims within a certain period; the claim amount is the specific amount of money actually paid out for claims; the historical policy details include more detailed policy-related content such as the start and end time of the policy and the scope of protection. These information are very important for analyzing the past insurance risk status of group customers.

[0083] For example, if the customer information shows that the company is mainly engaged in the manufacture of electronic products and is large in scale and in good operating condition, its industry label is determined as "electronic manufacturing" through relevant rules. The insurance policy and compensation information are determined based on the historical insurance data, and specific insurance policy and compensation related information such as insurance type and claim records are extracted and sorted out from a large number of historical insurance data. The customer information and historical insurance data of the group customers are input into the risk assessment module according to the requirements of the module, and the module is allowed to process these data according to its internally set algorithms and models, so as to obtain the risk assessment results corresponding to each industry label.

[0084] Through such a setting, this plan can comprehensively collect customer information and historical insurance data of group customers, and carefully sort out the key elements to determine industry labels and policy payment information, and then use the risk assessment module for analysis and processing. The beneficial effect of this is that it can deeply and accurately explore various situations of group customers at the insurance level, making the risk assessment results more scientific and reliable, providing detailed and accurate basis for subsequent insurance business development, risk control and related decision-making, and helping the insurance business to operate more reasonably and efficiently.

[0085] In one embodiment, optionally, the method further comprises:

[0086] Establish basic analysis model;

[0087] Using customer information and historical insurance data of group customers as sample data, respectively constructing a training set and a test set, after using the training set to train the basic analysis model, if the risk assessment result of the test set using the trained analysis model meets the preset accuracy rate, then determining that the analysis model training is completed;

[0088] Accordingly, based on the customer information and historical insurance data of the group customers, the risk assessment module is used to obtain the risk assessment results corresponding to each industry label, including:

[0089] The analysis model of the risk assessment module is used to analyze the customer information and historical insurance data of the group customers to obtain risk assessment results corresponding to various industry labels.

[0090] Among them, the basic analysis model can be an initial model framework whose functions need to be improved through data training. It is usually built based on a specific mathematical algorithm or machine learning algorithm. For example, it can be an infrastructure built based on a decision tree algorithm. It does not initially have the ability to accurately analyze the risks of group customers and requires subsequent training and optimization with data.

[0091] The training set can be a part of data samples selected from the customer information of group customers and historical insurance application data. These sample data will be used to train the basic analysis model so that it can learn the correlation rules between the features in the data and the risk assessment results. For example, customer information and corresponding insurance application data of a certain number of enterprises in different industries and of different scales are selected as the content of the training set.

[0092] The test set is also another part of data samples divided from the customer information of group customers and historical insurance application data. Its function is to test the accuracy and generalization ability of the model after the basic analysis model is trained by the training set, that is, to verify whether the model can accurately output the risk assessment results that meet the expectations on the unseen data. For example, relevant data of some enterprises are reserved to form the test set to test the performance of the trained model.

[0093] The preset accuracy rate can be a pre-set standard measurement value used to judge whether the trained analysis model has reached an acceptable level of accuracy, usually presented in the form of a percentage. For example, if the preset accuracy rate is set at 90%, when the proportion of the results obtained by the model for risk assessment of the test set that match the actual situation reaches this value, it is considered that the model training is qualified. It is an important reference index for measuring the training effect of the model.

[0094] In this solution, the data can be divided in the ratio of 8:2, with 80% as the training set and 20% as the test set, so as to reasonably arrange the data for the training and verification of the model.

[0095] The training of the model can be to input the data of the training set into the basic analysis model, so that the model continuously adjusts its internal parameters, weights, etc. according to various features in the data and the corresponding risk assessment results and other information, in order to learn how to accurately output the risk assessment results based on the input customer information and historical insurance application data.

[0096] After the training is completed, the customer information and historical insurance application data of group customers can be input into the analysis model of the trained risk assessment module, so that the model processes and calculates these data according to the rules and capabilities it has learned, and thus outputs the risk assessment results corresponding to each industry label.

[0097] This solution trains and tests by establishing a basic analysis model and using the customer information and historical insurance application data of group customers, and determines that the model training is completed according to a preset accuracy rate. Then, the trained model is used to analyze relevant data to obtain a risk assessment result. Its beneficial effect is that by means of the training mechanism of machine learning, the risk assessment model can better fit the actual business data situation, continuously optimize its own analysis ability, and thus improve the accuracy and reliability of the risk assessment result, providing strong technical support for accurately grasping customer risks and reasonably formulating insurance strategies in the insurance business.

[0098] In one embodiment, optionally, using the customer information of group customers and historical insurance application data as sample data includes:

[0099] Obtaining the basic data of the customer information of the group customers and historical insurance application data;

[0100] Performing data cleaning on the basic data to obtain cleaned result data; wherein, the data cleaning includes one or more of removing duplicate records, handling missing values, filling specific values, and handling outliers;

[0101] Using the cleaned result data as sample data.

[0102] Among them, the basic data can be the original content of the customer information of group customers and historical insurance application data obtained initially. These data may come from a wide range of sources, such as being obtained from the enterprise internal management system, the past insurance business record system, etc. However, in the original state, there may be various problems such as non-standard formats and uneven data quality. For example, there may be duplicate insurance application information in it, and there may be missing values in some key fields. It is the data source for subsequent processing.

[0103] The cleaned result data can be relatively high-quality data content obtained after the data cleaning process. After removing duplicate records, removing redundant information such as exactly the same duplicate insurance application records to avoid interfering with subsequent analysis, handling missing values, adopting appropriate filling strategies to supplement situations such as the enterprise scale field being empty, filling specific values, for example, filling corresponding values according to rules for some fields with default values being empty but having agreed rules, and handling outliers, such as correcting or removing values of claim amounts that deviate greatly from the normal range, the data becomes more accurate, complete, and standardized, and is suitable as sample data for subsequent construction of training sets, etc.

[0104] This solution collects customer information of group customers and original basic data of historical insurance data through corresponding data collection channels, interfaces or extraction from storage databases, and summarizes and integrates them together for further processing, such as obtaining customer information by connecting with the company's ERP system and exporting historical insurance data from the insurance business database.

[0105] This solution first obtains basic data, then carefully cleans it, and finally uses the cleaned data as sample data. The beneficial effect of this is that it can effectively improve data quality and remove those adverse factors in the original data that may interfere with analysis and affect the model training effect. Subsequent work based on these sample data, such as building training sets, training analysis models, and risk assessment, can be carried out on the basis of high-quality data, thereby further ensuring the accuracy and reliability of the entire insurance business-related analysis and decision-making.

[0106] In one embodiment, optionally, the risk assessment module further includes a claims data integration model, and the training process of the claims data integration model includes:

[0107] Acquire the claims-related data in the historical insurance data; wherein the claims-related data includes one or more of the claim amount, claim frequency, claim reason, claim time, policyholder information and insurance terms;

[0108] The initial claims data integration model is established, and the claims-related data is used for training to obtain a claims data integration model; wherein the claims data integration model is used to output at least one result of trend analysis, correlation analysis, risk scoring and predictive analysis.

[0109] Among them, the claims data integration model can be a model in the risk assessment module that is specifically used for comprehensive processing and analysis of claims-related data. It can explore the patterns and information hidden behind the claims data, and by integrating various claims-related data, it can output valuable analysis results to assist in the judgment of the risk status of group customers and insurance business decision-making.

[0110] The initial claims data integration model is also a basic model framework that has not been trained and does not yet have actual analysis capabilities. It is usually built based on specific data analysis or machine learning algorithms. For example, its initial architecture can be built based on a neural network algorithm, and it will need to be continuously trained using claims-related data to improve its functionality.

[0111] Trend analysis can be an analysis of the changes in claims-related data within a certain time range. For example, observing whether the claim amount has been increasing or decreasing year by year in the past few years, or what kind of fluctuation trend the claim frequency shows, etc. Through this analysis, the development trend of the claims aspect in the insurance business can be understood, which helps to discover potential risks or problems in advance.

[0112] Association analysis can be used to explore the mutual relationships among various elements of claims-related data. For example, analyzing whether there is a certain association between the claim reason and the policyholder information, or whether there is a connection between different settings of the underwriting terms and the claim frequency, etc. Through this analysis, the key factors affecting the claim situation and their internal connections can be discovered, which is convenient for optimizing insurance business terms and risk prevention and control strategies.

[0113] Risk scoring can assign a score representing the level of risk to group customers or specific insurance businesses, etc. based on claims-related data and pre-set scoring rules. For example, with a full score of 100, the higher the score, the greater the risk. This score can intuitively reflect the risk situation, which is convenient for risk comparison and decision-making.

[0114] Predictive analysis can be based on past claims-related data. Through certain data analysis and prediction algorithms, the possible future claim situations can be estimated. For example, predicting the approximate range of the claim amount and the level of claim frequency within a certain future time period, etc. It can help insurance companies make preparations for risk response and resource allocation in advance.

[0115] This solution constructs a claims data integration model and conducts targeted training on it, using the claims-related data in the historical insurance application data to improve the model's functions, enabling it to output results such as trend analysis, association analysis, risk scoring, and predictive analysis. Its beneficial effect lies in that it can deeply explore the value in the claims data, more comprehensively and accurately grasp the dynamic changes, internal connections, and future trends of the claims aspect in the insurance business, thereby providing strong and detailed basis for insurance companies in many aspects such as risk control, business optimization, resource allocation, and decision-making, effectively improving the scientificity and rationality of insurance business operations.

[0116] In one embodiment, optionally, after obtaining the claims-related data in the historical insurance application data, the method further includes:

[0117] Performing feature engineering processing on the claims-related data;

[0118] After using the claims-related data for training to obtain a claims data integration model, the method further includes:

[0119] Use cross-validation and / or A / B testing to evaluate the performance of the claim data integration model, adjust the parameters of the claim data integration model according to the results of the model performance evaluation, or replace the claim data integration model.

[0120] Among them, feature engineering processing can be a process of preprocessing data, extracting features, and constructing. For claim-related data, it is to unify data of different magnitudes into a reasonable range through data standardization. For example, scale the claim amount uniformly to the range of 0-1. Feature selection can select the most valuable features for model training and analysis from many fields of claim-related data. For example, determine which of the claim reason and claim frequency has a greater impact on the final result and select the more critical one to keep. Feature combination can be to combine some individual features to create new meaningful features, such as combining the claim time and claim amount to see if it can reflect seasonal claim patterns and other operations, making the data more suitable for the model to learn and process, and mining more valuable information.

[0121] Cross-validation is a common method for evaluating model performance. It divides the data set into multiple parts. For example, in the common k-fold cross-validation, assuming k = 5, the data is evenly divided into 5 parts, and one part is used as the test set in turn, and the remaining parts are used as the training set. Repeat this process multiple times and calculate the evaluation metrics each time, such as accuracy, mean squared error, etc. Finally, comprehensively evaluate the stability and generalization ability of the model on different data subsets based on these results to judge the overall performance of the model.

[0122] A / B testing can generally be used to compare the effects of different models or different versions of the same model. Usually, group A and group B are set. Group A uses the original model or the original parameter configuration, and group B uses the new model or the adjusted parameter configuration. Then, in the same test environment, the same test data is input to the two groups, and the output results are compared, such as the accuracy of the risk score, the fitting degree of the predictive analysis, etc. Determine which model or parameter configuration is better according to the comparison.

[0123] In this solution, when performing feature engineering processing, various data processing and feature construction methods mentioned above can be used to operate on claim-related data to make it more in line with the requirements of model training, optimize the feature representation of the data, and improve the efficiency and effect of the subsequent model in learning the internal laws of the data.

[0124] When evaluating the model performance, means such as cross-validation and / or A / B testing can be used to detect the performance of the claim data integration model under different data conditions and different configuration comparisons, and collect relevant evaluation index information such as accuracy and error, so as to judge whether the model meets the expected performance standards.

[0125] Based on the results of the model performance evaluation, the parameters within the claims data integration model, such as the weights and biases in the neural network model, or the node splitting threshold in the decision tree model, can be modified to improve the performance of the model so that it can more accurately output the desired results such as trend analysis and correlation analysis.

[0126] If after model performance evaluation it is found that the performance of the current claims data integration model is too poor and cannot be effectively improved by adjusting parameters, then choose to replace it with another more suitable model, such as changing from a linear regression-based model to a support vector machine-based model, to obtain better model performance and ensure the accuracy of subsequent insurance business analysis.

[0127] This solution performs feature engineering on claims-related data to make the data better adapt to model training and mine more valuable feature information, so as to use cross-validation and AB testing to evaluate model performance and adjust or replace the model accordingly. It can accurately control the quality and applicability of the model, so that the claims data integration model can continue to maintain a good performance status and output more reliable and accurate analysis results, thereby providing solid and accurate data support for risk assessment, decision-making and other links in the insurance business, and effectively improving the scientificity and effectiveness of the insurance business.

[0128] In one embodiment, optionally, the insurance business analysis module includes a pricing model;

[0129] Among them, the pricing model is used to determine the insurance pricing result of the current insurance business based on the group user portrait of the target industry label and the historical insurance record of the insured subject in the insured subject information.

[0130] Among them, the pricing model can be a key functional model in the insurance business analysis module. It mainly calculates and determines the price of insurance business based on specific input information, according to pre-set algorithms, rules and related risk assessment logic. In this scenario, it is based on key materials such as the group user portrait corresponding to the target industry label and the historical insurance records of the insured entity to output reasonable insurance pricing results. For example, a model built based on actuarial principles combined with machine learning algorithms quantifies different influencing factors and relates them to the final pricing.

[0131] The insurance pricing result is the specific insurance premium price value or price range given for the current insurance business after the analysis and calculation of the pricing model. It comprehensively considers the characteristics of the industry in which the insured entity is located and its past insurance history. It is the key conclusion finally determined to guide the actual charging of this insurance business. For example, the pricing result of a company's property insurance is an annual premium of 10,000 yuan.

[0132] In this solution, the pricing model is based on the input information such as the group user portrait of the target industry label received and the historical insurance records of the insured entity. According to its internal pricing logic, calculation rules, etc., after a series of operations such as analysis, processing, and calculation, the accurate and reasonable insurance pricing result of the current insurance business is finally obtained.

[0133] In this solution, the pricing model included in the insurance business analysis module determines the insurance pricing result by combining the group user portrait of the target industry label with the historical insurance records of the insured entity. Its beneficial effect is that it can fully consider various factors such as the industry characteristics and past insurance situations of the insured entity, making the pricing more scientific, reasonable, and targeted, avoiding the blindness of pricing, and not only ensuring the reasonable income of the insurance company but also enabling the insured customers to feel that the pricing conforms to their actual situations.

[0134] Example 2

[0135] In order to enable those skilled in the art to understand this solution more clearly, this application also provides a preferred embodiment.

[0136] In this solution, first, a model is established using big data. From the historical insurance data of corporate group customers, data such as the insurance preferences, loss ratios, and claim settlement habits of individual customers are analyzed, key data is extracted, and group characteristics such as industry and scale are added to establish an analysis model.

[0137] Data input: Information is collected from the historical insurance data of corporate group customers, including but not limited to the insurance preferences of individual customers (such as insurance types, insurance amounts), loss ratios, claim settlement records, claim settlement frequencies, claim settlement amounts, and historical policy details. At the same time, external data such as industry classifications, enterprise scales, operating conditions, and geographical locations are integrated as features.

[0138] Specifically, first, the collected data can be cleaned, including: removing duplicate records, handling missing values, filling specific values, and handling outliers; among them,

[0139] Removing duplicate records: The unique identifiers of the data (such as customer IDs, order numbers) are mapped to the hash table through the hash function, and duplicates are quickly identified and removed.

[0140] Handling missing values: By the deletion method: directly deleting the rows or columns containing missing values, which is applicable to the situation where there are many missing values or the data volume is large and the missing values have little impact on the analysis. Filling method: Mean / median / mode filling: applicable to numerical data, filling the missing values with the mean, median, or mode of this column.

[0141] Specific value filling: According to business logic, select specific values for filling. For example, fill a certain category as "unknown". Model prediction filling: Use machine learning models (such as KNN, decision tree, random forest) to predict missing values, and make predictions based on the complete information of other columns. Imputation method: For time series data, use the previous or next value for forward or backward filling (forward filling or backward filling).

[0142] Outlier handling: Through statistical methods: Interquartile Range (IQR) rule: Identify data outside the range of Q1 - 1.5 times IQR or Q3 + 1.5 times IQR as outliers. Z-score standardization: Calculate the standard deviation distance of data points from the mean. Usually, data exceeding 3 times the standard deviation is regarded as an outlier. Box plot: Intuitively display the data distribution. Outliers are usually shown as outlier points and can be directly identified and processed. Cluster analysis: Such as DBSCAN (density-based clustering algorithm), etc., can identify low-density isolated points as outliers. Machine learning methods: Use algorithms such as Isolation Forest, LocalOutlier Factor (LOF), etc., which are specifically used for outlier detection.

[0143] The specific implementation steps are as follows:

[0144] Import data: Use libraries such as Pandas to read data;

[0145] Remove duplicates: df.drop_duplicates(inplace = True);

[0146] Check for missing values: Use df.isnull().sum() to view the missing situation of each column;

[0147] Handle missing values: Select a suitable method, such as using the median to fill with df.fillna(df['column'].median());

[0148] Identify outliers: Calculate the Z-score or use the IQR rule to mark outliers;

[0149] Handle outliers: Decide to delete, replace, or correct outliers according to business logic;

[0150] Save the cleaned data: df.to_csv('cleaned_data.csv', index = False).

[0151] Through the model, focus on the underwriting and claim analysis of corporate customers by industry, and discover the diseases with high incidence and high claim liability types in specific industries;

[0152] After that, build an industry group customer profile. Based on the historical underwriting and claim settlement situations of the group as a whole, analyze indicators such as its risk exposure count, claim occurrence rate, and claim settlement habits to construct a comprehensive risk assessment report;

[0153] Establish a group customer management system. To better evaluate the underwriting and claim settlement situations of large group customers, a dedicated group customer management system can be established. This system can centrally manage all the insured entities, business data, and claim settlement information related to the group. By integrating this information, a more comprehensive understanding of the customer's business situation, risk status, claim settlement history, etc. can be obtained;

[0154] Introduce data analysis and mining tools. Use advanced data analysis and mining tools, such as big data analysis, artificial intelligence, etc., to deeply analyze the underwriting and claim settlement situations of each insured entity and the entire group. These tools can help us discover the laws and trends behind the data and provide more accurate support for business decisions;

[0155] Establish a claim settlement data analysis model. By establishing a claim settlement data analysis model, in-depth analysis of the claim settlement data of each insured entity can be carried out. These models can help us understand the claim settlement situations and risk statuses of each insured entity and provide more accurate support for refined pricing and large and medium-sized order businesses;

[0156] Composition of the claim settlement data analysis model:

[0157] Data collection and preprocessing:

[0158] First of all, various types of data related to claim settlement need to be collected, including but not limited to claim amount, claim frequency, claim reason, claim time, insured information (such as age, occupation, health status), underwriting terms, historical claim settlement records, etc.

[0159] Feature engineering: Extract useful features from the original data, which may include calculating certain statistics (such as average claim amount, standard deviation), creating derivative variables (such as classifying according to claim reasons), data standardization or normalization, etc.

[0160] Model selection: Select suitable statistical models or machine learning algorithms, such as logistic regression, random forest, gradient boosting tree, neural network, etc., for predicting claim settlement risks or analyzing claim settlement patterns.

[0161] Model training: Use historical data to train the selected model, and iteratively optimize the model parameters to make it fit the data as accurately as possible.

[0162] Verification and adjustment: Evaluate the model performance through methods such as cross-validation, A / B testing, etc., and adjust the model parameters or replace the model according to the evaluation results to improve the prediction accuracy.

[0163] Explanation and Application: Interpret the model results to understand which factors have the greatest impact on claim risk, and accordingly formulate risk management strategies or pricing policies.

[0164] Among them, the specific types of data can include the following:

[0165] Claim data: claim amount, claim date, claim type, and processing time, etc.;

[0166] Policyholder information: age, gender, occupation, health status, and past medical history, etc.;

[0167] Policy details: type of insurance, scope of coverage, premium, deductible, and insurance term, etc.;

[0168] Historical claim records: past claim frequency and total claim amount for an individual or a group, etc.;

[0169] External data: macroeconomic indicators, industry-specific data, such as accident rates for auto insurance, etc.;

[0170] For the principle of the modeling formula, the following methods can be adopted:

[0171] Taking logistic regression as an example, it is often used for binary classification problems, such as determining whether a certain insured entity has a high claim risk. The logistic regression model is based on the following formula:

[0172] Steps for in-depth analysis of claim data for each insured entity;

[0173] Data segmentation: Group the data by insured entity to ensure that each group of data contains all the claim records of a single entity.

[0174] Outlier detection: Identify and handle extreme or unreasonable data points to ensure the accuracy of the analysis results.

[0175] Trend analysis: Analyze the changing trends of claim frequency and amount over time for each insured entity.

[0176] Association analysis: Explore the relationships between different variables, such as the relationship between claim amount and the age and occupation of the policyholder.

[0177] Risk scoring: Use the model to calculate a risk score for each insured entity to reflect its potential claim risk.

[0178] Predictive analysis: Predict future claim trends based on current data to help optimize resource allocation and pricing strategies.

[0179] Customized recommendations: Provide personalized risk management recommendations or product customization solutions for each insured entity based on the analysis results.

[0180] In this solution, pricing support tools can be introduced. To achieve refined pricing for large and medium-sized order business, pricing support tools such as pricing models and rate calculation tools can be introduced. These tools can provide personalized pricing solutions based on the risk profile and business characteristics of each insured entity. These solutions can provide more accurate support for business decisions and improve the revenue level of the business; and specifically include the following processing procedures:

[0181] Data integration and cleaning:

[0182] Data collection: First, collect relevant data from various sources (such as CRM systems, historical policy databases, third-party data providers), including but not limited to the basic information of the insured entity (such as enterprise scale, industry classification, financial status), historical claim records, insurance preferences, historical loss ratios, claim settlement habits, etc.

[0183] Data cleaning: Perform data deduplication, missing value handling (such as filling with mean, median or model prediction), outlier handling (such as removing or transforming with Z-score method), and data type conversion to ensure data quality.

[0184] Feature engineering:

[0185] Feature selection: Based on business understanding, select features that have a direct impact on pricing decisions, such as industry risk level, historical loss ratio, company credit score, business scale, etc.

[0186] Feature construction: Create new derived variables, such as historical claim frequency, average claim amount ratio, business growth trend, etc., to enhance the predictive ability of the model.

[0187] Model building:

[0188] Model selection: Select a suitable pricing model according to business needs, such as linear regression, logistic regression (suitable for continuous premium prediction), decision tree, random forest, gradient boosting model (XG Boosting), neural network, etc.

[0189] Model training: Use the cleaned and constructed dataset for model training, divide the training set and validation set to evaluate the model performance, and adjust the model parameters to optimize the model performance.

[0190] Verification and optimization:

[0191] Model verification: Verify the model on the test set and evaluate the prediction accuracy of the model through indicators such as KPIE, RMSE, R 2 and other metrics.

[0192] Model calibration: If there is a deviation in model prediction, perform calibration, such as adjusting factors to ensure that the premium matches the risk.

[0193] Sensitivity Analysis: Conduct sensitivity tests to understand the impact of different parameters on pricing results and ensure the robustness of the model.

[0194] Application and Decision Support:

[0195] Personalized Pricing: Generate personalized rates and pricing plans for each insured entity based on the output of the pricing model, considering their risk characteristics and business features.

[0196] Decision Support Report: Generate a detailed pricing report, including pricing basis, risk analysis, sensitivity analysis, etc., to provide comprehensive support for the decision-making level.

[0197] Dynamic Adjustment: Continuously monitor the performance of the pricing model based on market feedback and new data, and timely adjust the pricing model parameters to maintain the accuracy of the pricing strategy.

[0198] In addition, this solution introduces customer classification management tools. To better manage group customers in a personalized manner, customer classification management tools can be introduced. These tools can automatically classify customers into different categories based on information such as customers' historical data, business characteristics, and risk status. Providing personalized services and support for different categories of customers can improve customer satisfaction and loyalty;

[0199] In addition, this solution can enhance the detail level of existing support tools. For existing support tools, their detail level can be improved by adding detailed claim settlement details and analysis functions. This allows users to more intuitively understand the claim settlement situation and risk status of each insured entity. At the same time, personalized claim settlement suggestions and service support can also be provided for each insured entity;

[0200] To better implement these solutions, relevant personnel need to be trained and supported. The training content includes the use of the new system, the application of new tools, data analysis techniques, etc. At the same time, personnel support needs to be provided, such as setting up dedicated account managers or technical support teams to provide timely help and support to customers;

[0201] Finally, optimize the insurance plan and product pricing through the customer analysis and evaluation report. Thus, a complete set of precise pricing solutions for employee welfare group insurance for the industry and specific large group customers is established to solve the problems of inaccurate selection of insured responsibilities for customer employee welfare insurance and inaccurate pricing by the underwriter.

[0202] Under the background of the rapid growth of employee welfare group insurance in enterprises, this solution provides a precise industry and customer underwriting analysis tool, improves the accuracy of product pricing, reduces the comprehensive loss ratio of policies, and effectively improves the underwriting quality of large customers;

[0203] Overall short-term insurance portfolio management, optimize short-term insurance pricing support;

[0204] Build a product claim assessment model relying on big data analysis;

[0205] Implement group management for key customers and optimize the pricing strategies for major customers;

[0206] Predict the company's expense creation situation with the help of machine learning and artificial intelligence, and allocate resources reasonably;

[0207] Overall management of decentralized combined short-term insurance business, improve the level of intelligent analysis, effectively monitor the short-term insurance business benefits of the company, and boost the scientific and intelligent decision-making of short-term insurance operations.

[0208] This solution provides a precise group policy issuance method for industry customers based on big data modeling and analysis technology. This method effectively utilizes big data modeling and analysis technology to build an experience knowledge base for group customers in different industries, and on the basis of combining the historical claim situations of insured customers, establish a set of precise group policy issuance solutions for industry customers; with the help of machine learning and artificial intelligence, the system can predict the company's expense creation situation through learning historical data, and with the help of artificial intelligence technology, allocate limited resources to more effective places; overall management of decentralized combined short-term insurance business, improve the level of intelligent analysis, effectively monitor the short-term insurance business benefits of the company, boost the scientific and intelligent decision-making of short-term insurance operations, and provide professional business data analysis tools.

[0209] Its beneficial effects include but are not limited to the following points:

[0210] 1. This method used to manually calculate the insured's risk exposure and risk incidence by regularly extracting the list of insured persons and the claim details list; now it can evaluate claim changes and risk levels in real time through the system, greatly improving accuracy and timeliness, and providing timely and necessary decision support for the market competition of large order business;

[0211] 2. This method includes classifying group customers and formulating different pricing strategies according to different types of customers. This classification method can be based on factors such as the customer's business scale, business type, risk status, etc., or can be based on factors such as the customer's consumption habits, credit rating, etc. This pricing strategy can include various pricing strategies such as risk assessment-based pricing, market condition-based pricing, and customer loyalty-based pricing.

[0212] 3. This method uses big data technology to collect and analyze the historical claim data of group customers, and formulates different pricing strategies according to the customer's claim situation and risk status. This method can include cleaning, sorting, analyzing, and mining historical claim data to obtain more accurate and comprehensive customer claim information. This pricing strategy can include various pricing strategies such as loss ratio-based pricing, claim time-based pricing, and claim amount-based pricing.

[0213] 4. This method utilizes data mining and machine learning technologies to collect and analyze the behavioral data of group customers, including factors such as customer consumption behavior, credit ratings, and behavior habits, and formulates different pricing strategies based on the behavioral characteristics of customers. This method may include classifying customer behavior using clustering algorithms and performing association analysis on customer behavior using association rule mining algorithms, etc. The pricing strategy may include various pricing strategies such as pricing based on customer behavior, pricing based on customer preferences, and pricing based on market competition.

[0214] 5. This method utilizes a risk assessment model to assess the risks of group customers and formulates different pricing strategies according to the risk status of customers. The risk assessment model may include evaluating multiple factors such as the business scale, business type, operating conditions, and financial conditions of customers to obtain a more accurate and comprehensive customer risk status. The pricing strategy may include various pricing strategies such as pricing based on the risk assessment results, pricing based on risk probabilities, and pricing based on risk management.

[0215] In addition, a claim combination runoff triangle analysis model can be used. Based on the historical product sales and claim data over the years, a runoff triangle analysis model for insurance types is constructed. For the premium, it is no longer simply calculated using the received premium, but the earned premium is calculated on a daily basis using the 365-day method. Combining with the claim payout rhythm analysis model, two empirical variables, namely the delay report time and the time required for case settlement, are generated to detail the entire claim process of the product policy, thereby estimating the number of claim cases and claim amounts in the entire life cycle of the product, scientifically calculating the outstanding claim reserve, optimizing the business development rhythm, and balancing the synchronous development of scale and efficiency.

[0216] Embodiment 3

[0217] Figure 2 is a schematic structural diagram of a policy processing device for group customers provided in Embodiment 3 of the present application. The device is configured in a policy management system, and the policy management system is deployed in an electronic device; as Figure 3 shown, the device includes:

[0218] A risk assessment module 210, configured to obtain the industry labels and historical insurance application data of group customers, and determine the risk assessment results corresponding to each industry label according to the industry labels and the historical insurance application data;

[0219] A portrait construction module 220, configured to generate group customer portraits corresponding to each industry label based on the risk assessment results corresponding to each industry label;

[0220] The industry matching module 230 is configured to obtain the applicant information when receiving an insurance application service, match the applicant information with the group customer portraits corresponding to each industry label that has been obtained, and obtain the target industry label corresponding to the applicant information;

[0221] The insurance application service analysis module 240 is configured to determine the processing result of the insurance application service according to the group user portrait of the target industry label and the historical insurance records of the applicant in the applicant information.

[0222] This device can execute the policy processing method for group customers provided in the above embodiments, and has corresponding functional units and beneficial effects. Details are not described herein again.

[0223] Embodiment 4

[0224] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0225] Therefore, the present application also proposes a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the method described in any one of the embodiments of the present application.

[0226] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the specified functions in one process Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0227] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means realizes the functions in one process Figure 1 one process or multiple processes and / or blocksFigure 1 The functions specified in one or more boxes.

[0228] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide for implementing the steps of the functions specified in one or more processes and / or boxes Figure 1 One or more processes and / or boxes Figure 1 The steps of the functions specified in one or more boxes.

[0229] Furthermore, Figure 3 FIG. [X] is a schematic structural diagram of an electronic device provided in Embodiment 4 of the present application. As Figure 3 shown, the present application also proposes an electronic device (or computing device), including a processor 11, a memory 12, and a computer program stored on the memory 12 and executable on the processor 11. When the processor 11 executes the computer program, the method described in any embodiment of the present application is implemented.

[0230] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and a memory. The memory may include non-permanent memory in the computer-readable medium, random access memory (RAM), and / or non-volatile memory in the form of, for example, read-only memory (ROM) or flash memory (flash RAM). The memory is an example of a computer-readable medium. Computer-readable media include both permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology for storing information. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transitory media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0231] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, commodity or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, commodity or device. Without further limitation, an element defined by the phrase "comprising an..." does not exclude the presence of additional identical elements in the process, method, commodity or device comprising said element.

[0232] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A method for processing insurance policies for group customers, characterized in that: The method is performed by an insurance policy management system, and the insurance policy management system is deployed in an electronic device; the method includes: Obtaining industry labels and historical insurance data of group customers, inputting the industry labels and historical insurance data into a risk assessment module, and obtaining risk assessment results corresponding to each industry label; Based on the risk assessment results corresponding to the various industry labels, the portrait construction module is called to obtain the group customer portraits corresponding to the various industry labels; When receiving the insurance business, the insured subject information is obtained, and the insured subject information is input into the industry matching module to obtain the target industry label corresponding to the insured subject information; The group user portrait of the target industry label and the historical insurance records of the insured subject in the insured subject information are input into the insurance business analysis module to obtain the processing result of the insurance business.

2. The method according to claim 1, characterized in that Obtain the industry labels and historical insurance data of the group customers, input the industry labels and historical insurance data into the risk assessment module, and obtain the risk assessment results corresponding to each industry label, including: Acquire customer information of group customers, and determine the industry label of the group customers according to the customer information; wherein the customer information includes one or more of industry classification, enterprise scale, operating status and geographical location; Acquire the historical insurance data of the group customers, and determine the insurance policy and compensation information according to the historical insurance data; wherein the historical insurance data includes one or more of the insurance type, insurance amount, claim record, compensation ratio, claim frequency, claim amount and historical insurance policy details; Based on the customer information and historical insurance data of the group customers, the risk assessment module is used to obtain risk assessment results corresponding to various industry labels.

3. The method according to claim 2, characterized in that The method further comprises: Establish basic analysis model; Using customer information and historical insurance data of group customers as sample data, respectively constructing a training set and a test set, after using the training set to train the basic analysis model, if the risk assessment result of the test set using the trained analysis model meets the preset accuracy rate, then determining that the analysis model training is completed; Accordingly, based on the customer information and historical insurance data of the group customers, the risk assessment module is used to obtain the risk assessment results corresponding to each industry label, including: The analysis model of the risk assessment module is used to analyze the customer information and historical insurance data of the group customers to obtain risk assessment results corresponding to various industry labels.

4. The method according to claim 3, characterized in that The customer information and historical insurance data of group customers are used as sample data, including: Obtaining basic data on customer information and historical insurance data of the group's customers; Performing data cleaning on the basic data to obtain cleaning result data; wherein the data cleaning includes: removing duplicate records, processing missing values, filling specific values, and processing abnormal values. The cleaning result data is used as sample data.

5. The method according to claim 1, characterized in that The risk assessment module also includes a claims data integration model. The training process of the claims data integration model includes: Acquire the claims-related data in the historical insurance data; wherein the claims-related data includes one or more of the claim amount, claim frequency, claim reason, claim time, policyholder information and insurance terms; The initial claims data integration model is established, and the claims-related data is used for training to obtain a claims data integration model; wherein the claims data integration model is used to output at least one result of trend analysis, correlation analysis, risk scoring and predictive analysis.

6. The method according to claim 5, characterized in that After obtaining the claims-related data in the historical insurance data, the method further includes: Performing feature engineering on the claims-related data; After using the claims-related data for training to obtain a claims data integration model, the method further includes: Use cross-validation and / or AB testing to evaluate the model performance of the claims data integration model, adjust the claims data integration model parameters according to the model performance evaluation results, or replace the claims data integration model.

7. The method according to claim 1, characterized in that The insurance business analysis module includes a pricing model; Among them, the pricing model is used to determine the insurance pricing result of the current insurance business based on the group user portrait of the target industry label and the historical insurance record of the insured subject in the insured subject information.

8. A policy processing device for group customers, characterized in that: The device is configured in an insurance policy management system, and the insurance policy management system is deployed in an electronic device; the device includes: The risk assessment module is used to obtain the industry labels and historical insurance data of the group customers, and determine the risk assessment results corresponding to each industry label based on the industry labels and the historical insurance data; A portrait construction module, used to generate group customer portraits corresponding to each industry label based on the risk assessment results corresponding to each industry label; The industry matching module is used to obtain the insured subject information when receiving the insurance business, match the insured subject information with the group customer portrait corresponding to each industry label that has been obtained, and obtain the target industry label corresponding to the insured subject information; The insurance business analysis module is used to determine the processing result of the insurance business based on the group user portrait of the target industry label and the historical insurance record of the insured subject in the insured subject information.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

Images