A data analysis method, system and storage medium for automobile insurance business

Abstract

The present invention discloses a data analysis method, system and storage medium for automobile insurance business, which relates to the field of data analysis technology, and includes the following steps: obtaining and integrating original automobile insurance business data from different channels and sources from the internal system of an insurance company, and performing data processing to integrate and obtain an insurance business sequence table; based on the original automobile insurance business data in the insurance business sequence table, identifying and analyzing business risk indicators and compensation risk indicators. The present invention constructs a diversified assessment model to conduct a comprehensive assessment of different types of risk events, greatly enhancing the comprehensiveness and accuracy of risk assessment, and can more accurately capture risk signals, warn of potential risks in advance, and provide a more reliable basis for decision-making. At the same time, it promotes the continuous improvement and optimization of the risk assessment system and promotes the intelligent development of automobile insurance business.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, specifically to a data analysis method, system, and storage medium for automobile insurance business. Background Technology

[0002] With the continuous increase in car ownership, the auto insurance market continues to expand. This trend has prompted the insurance industry to adopt more refined management methods to cope with the growing business demands. With the development of internet technology, more and more consumers are choosing to purchase auto insurance products online, and consumers' demands for auto insurance products are becoming more diversified. In order to meet the needs of different consumers, the insurance industry needs to use data analysis to gain a deeper understanding of customer needs and preferences, thereby developing more targeted insurance products.

[0003] In existing technologies, auto insurance business involves a wide variety of data. When evaluating auto insurance business, only a few key indicators are often focused on, leading to one-sided evaluation results that fail to fully reflect the business situation. Furthermore, auto insurance business faces various risks, and insufficient attention to these risks during the evaluation process often results in inaccurate risk assessments. Therefore, how to establish a diversified evaluation system that comprehensively considers multiple factors for business and risk assessment is the problem we need to solve. To this end, we propose a data analysis method, system, and storage medium for auto insurance business. Summary of the Invention

[0004] The purpose of this invention is to provide a data analysis method, system, and storage medium for automobile insurance business, in order to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0006] Firstly, a data analysis method for auto insurance business includes the following steps:

[0007] Step 1: Obtain and integrate raw data of auto insurance business from different channels and sources from the insurance company's internal system, process the data, and integrate it to obtain the insurance business sequence list;

[0008] Step 2: Based on the original data of auto insurance business in the insurance business sequence table, identify and analyze the business risk indicators and claims risk indicators. Based on the data analysis results, extract the sub-indicator characteristics of the business risk indicators and claims risk indicators. Among them, the business risk indicators are used to analyze the historical trends and current status of auto insurance business, understand the overall situation and risks of business development, and the claims risk indicators are used for monitoring and early warning to promptly identify potential risk points.

[0009] Step 3: Combining the original data of the auto insurance business, conduct multi-dimensional analysis of business risk indicators and claims risk indicators, construct business risk prediction models and claims risk prediction models respectively, and obtain business risk assessment index and claims risk assessment index.

[0010] Step 4: Based on business characteristics and risk assessment needs, comprehensively analyze business risk indicators and claims risk indicators, establish a diversified assessment model, and combine the business risk assessment index and claims risk assessment index to obtain diversified risk prediction coefficients, and comprehensively assess the impact of different risk factors on auto insurance business.

[0011] Step 5: Based on the multivariate risk prediction coefficients and the original data of the auto insurance business, classify the risk warning levels according to the recorded risk events, and set corresponding risk prediction thresholds for the risk warning levels.

[0012] Step 6: Based on the established diversified assessment model, build a real-time monitoring system to monitor new data streams and identify risk events, and generate risk assessment reports, including risk factor analysis, interpretation of prediction results, and recommended measures to optimize the implementation strategy.

[0013] A further improvement to the technical solution of this invention lies in the following: In step 1, the process of obtaining the insurance business sequence table is as follows:

[0014] Step 101: Based on the insurance company's internal CRM system, claims system, and financial system, and in conjunction with the bank's credit system, obtain data access permissions to determine the categories of raw data for auto insurance business. Raw data for auto insurance business includes customer information, vehicle information, policy details, claims records, claims process, customer service records, and credit records. Customer information includes name, contact information, and address; vehicle information includes license plate number, vehicle model, purchase date, and vehicle value; policy details include insurance type, coverage amount, insurance period, and premium; claims records include claims amount, claims time, and claims reason; claims process includes reporting time, investigation record, and damage assessment result; customer service records include consultation content and complaint handling status; and credit records include loan amount, repayment history, and credit score. The CRM system is the Customer Relationship Management System.

[0015] Step 102: Use an ETL tool to extract the required categories of original auto insurance business data from each system and back up the data to prevent data loss or damage.

[0016] Step 103: Clean the collected raw data of auto insurance business, including removing duplicate records, handling missing values, and correcting erroneous data to improve data quality, and convert data from different sources into a unified format and standard for subsequent processing and analysis, and integrate the cleaned and transformed data into a unified data warehouse.

[0017] Step 104: Construct an insurance business sequence table. Fill the integrated data according to the structure of the sequence table, and verify the filled sequence table by comparing it with the original data of the auto insurance business to ensure the accuracy and completeness of the data.

[0018] A further improvement to the technical solution of this invention lies in the following: In step 2, the extraction process of the features of the business risk indicators and the compensation risk indicators sub-indicators is as follows:

[0019] Step 201: Review the insurance business sequence table to clarify its data structure and content;

[0020] Step 202: Define business risks and identify business risk indicators based on the data in the insurance business sequence table. These business risk indicators include business growth rate, premium growth rate, customer retention rate, renewal rate, and insurance penetration rate. Analyze the historical trends of the business risk indicators, further refine the business risk indicators, extract their sub-indicator characteristics, and conduct in-depth analysis of the changing trends of the business risk indicators.

[0021] Step 203: The sub-indicators of business growth rate are the annual policy quantity growth rate and the annual premium income growth rate; the sub-indicators of premium growth rate are the new product sales growth rate and the supplementary protection sales growth rate; the sub-indicators of customer retention rate are the annual customer retention rate and the high-value customer retention rate; the sub-indicators of renewal rate are the renewal rate of different product lines and the average number of policies per renewing customer; the sub-indicators of insurance penetration rate are the regional insurance coverage rate and the insurance participation rate; the annual policy quantity growth rate is used to compare the growth rate of the number of new policies signed each year; the annual premium income growth rate is used to assess the growth of premium income each year; and the new product sales growth rate is used to assess the market acceptance of newly launched insurance products. The following metrics are used to analyze insurance coverage and growth potential: Sales growth rate of additional coverage products; annual customer retention rate; high-value customer retention rate; renewal rate of different product lines; average number of policies per renewing customer; regional insurance coverage rate; and insurance participation rate.

[0022] Step 204: Define claims risk and identify claims risk indicators based on the data in the insurance business sequence table. The claims risk indicators include the loss ratio, claims cycle, and average claim amount per case. Establish a claims risk monitoring and early warning mechanism, further refine the claims risk indicators, extract the characteristics of their sub-indicators, and conduct in-depth analysis of the potential risk points of the claims risk indicators.

[0023] Step 205: The sub-indicators of the loss ratio are the basic loss ratio and the loss ratio of different product lines; the sub-indicators of the claims cycle are the average claims processing time and the processing time of different types of claims; the sub-indicators of the average claim amount per case are the average claim amount per insurance type and the maximum difference in claim amount per insurance type. The basic loss ratio is used to calculate the ratio of total claims to total premium income; the loss ratio of different product lines is used to compare the loss ratios of different insurance products and identify high-risk product lines; the average claims processing time is used to analyze the average time from claim application to claim completion and to evaluate claims processing efficiency; the processing time of different types of claims is used to compare the average processing time of different types of claims and to identify potential processing bottlenecks and improvement points; the average claim amount per insurance type is used to compare the average claim amount of different types of claims and to identify the characteristics of high-claim cases; the maximum difference in claim amount per insurance type is used to analyze the distribution of claim amounts and to identify potential high-risk cases.

[0024] A further improvement to the technical solution of this invention lies in the following: In step 3, the process of obtaining the business risk assessment index and the compensation risk assessment index is as follows:

[0025] Step 301: Extract feature data of business risk indicators and claims risk indicators from the original data of automobile insurance business in the insurance business sequence table to obtain business risk feature dataset and claims risk feature dataset. Divide the business risk feature dataset and claims risk feature dataset into training set and test set respectively.

[0026] Step 302: Based on the business risk indicators, extract their sub-indicator features, use the training set data of the business risk feature dataset to train the model, build a business risk prediction model, and use the test set of the business risk feature dataset to evaluate and test the business risk prediction model in order to analyze the risk status of the auto insurance business.

[0027] Step 303: Based on the payout risk index, extract its sub-index features, use the training set data of the payout risk feature dataset to train the model, build a payout risk prediction model, and use the test set of the payout risk feature dataset to evaluate and test the payout risk prediction model in order to predict payout risk.

[0028] Step 304: Analyze the business risk prediction model and the compensation risk prediction model to obtain the business risk assessment index and the compensation risk assessment index. The business risk assessment index is used to quantify the level of business risk, and the compensation risk assessment index is used to quantify the level of compensation risk.

[0029] Step 305: Conduct risk assessment and monitoring based on the results of the business risk assessment index and the compensation risk assessment index.

[0030] A further improvement to the technical solution of this invention lies in the fact that the calculation formula for the business risk assessment index is:

[0031]

[0032] Among them, BI is the Business Risk Assessment Index, and G... i It is the business growth rate at the i-th time point. σ(R) is the average of the business growth rate, R is the customer retention rate, σ(R) is the standard deviation of the customer retention rate, max(R) is the maximum value of the customer retention rate, min(R) is the minimum value of the customer retention rate, PR is the insurance penetration rate, which measures the degree of popularity of insurance products in a specific market, n is the number of time points, and μ(R) is the average of the customer retention rate.

[0033] The formula for calculating the compensation risk assessment index is as follows:

[0034]

[0035] max(V)=max(V1,V2,…,V K );

[0036] CI stands for Claims Risk Assessment Index. j P is the payout amount for the j-th claim. j Is with C j The corresponding premium income, m is the number of claims, used as the exponent in the square root, CT is the claims processing time, which is the average time from claim application to payment completion, max(CT) is the maximum claim processing time, and min(CT) is the minimum claim processing time. V represents the average payout per case, and max(V) represents the maximum variance of the average payout per case, indicating the volatility of the payout. k It is the variance of the average claim amount, K is the number of time periods, and o k It refers to the number of cases within a given time period, C. jk It is the payout amount for the j-th claim in the k-th time period, μ(C k () is the average payout amount over the time period.

[0037] A further improvement to the technical solution of this invention lies in the following: In step 4, the process of obtaining the multivariate risk prediction coefficients is as follows:

[0038] Step 401: Extract feature datasets of business risk and claims risk from the raw data of auto insurance business, and divide the extracted feature datasets into training set and test set;

[0039] Step 402: Using the training set data and the characteristics of business risk indicators and compensation risk indicators, a diversified assessment model is established using a multiple linear regression model, and the diversified assessment model is tested and analyzed using test set data.

[0040] Step 403: Conduct a comprehensive analysis of the diversified assessment model, and combine the business risk assessment index and the claims risk assessment index to obtain the diversified risk prediction coefficient. Use the diversified risk prediction coefficient and the diversified assessment model to conduct a risk assessment of the current business and analyze the comprehensive impact of different risk factors on the auto insurance business.

[0041] A further improvement to the technical solution of this invention lies in that: the calculation formula for the multivariate risk prediction coefficient is:

[0042]

[0043] Where MRC is the multivariate risk prediction coefficient, BI l It is the l-th business risk assessment index, CI l It is the l-th compensation risk assessment index, and v is the total number of features of business risk indicators and compensation risk indicators. It is the average business growth rate, used in limit functions to represent the long-term growth trend.

[0044] A further improvement to the technical solution of this invention lies in the following: In step 5, the process of setting the risk prediction threshold for the risk warning level is as follows:

[0045] Step 501: Based on the collected original data of the auto insurance business, analyze the recorded risk events, determine the type and severity of the risk events, and the types of risk events include claims events, fraudulent behavior, and high-risk driving behavior;

[0046] Step 502: Based on the determined types of risk events, combined with the distribution of multivariate risk prediction coefficients and the severity of risk events, risk warning levels are divided into low-risk warning level, medium-risk warning level and high-risk warning level.

[0047] Step 503: Based on the risk prediction value of the multivariate risk prediction coefficient, set a corresponding risk prediction threshold for each risk warning level;

[0048] Multiple risk warning levels correspond to multiple risk prediction thresholds, wherein the risk prediction thresholds include an upper limit threshold and a lower limit threshold;

[0049] The multiple risk warning levels and the multiple risk prediction thresholds satisfy the following relationship:

[0050] Low risk warning level MRC≥MRC zy A lower risk prediction value indicates that the policyholder has a lower risk profile and is less likely to experience claims, fraud, or high-risk driving behavior.

[0051] Medium risk warning level MRC gy ≤MRC <MRC zy The risk prediction value is at a moderate level, indicating that the policyholder faces certain risks and their claims and driving behavior need to be monitored.

[0052] High-risk warning level 0 <MRC<MRC gy A high risk prediction value indicates that the policyholder faces a higher risk of claims, fraud, or high-risk driving behavior.

[0053] Where MRC represents the multivariate risk prediction coefficient, MRC zy MRC sets the lower threshold for low-risk warning levels and the upper threshold for medium-risk warning levels. gy These are the lower threshold corresponding to the medium-risk warning level and the upper threshold corresponding to the high-risk warning level.

[0054] A further improvement to the technical solution of this invention lies in the following: In step 6, the process of generating the risk assessment report is as follows:

[0055] Step 601: Establish a data pipeline to collect real-time data streams from the auto insurance business, including policy information, claim records, driving behavior data, etc. Ensuring the real-time nature and integrity of the data is the foundation of system construction.

[0056] Step 602: Integrate the diversified assessment model into the monitoring system and use the diversified assessment model to perform real-time analysis of the data flow of the auto insurance business to identify potential risk events.

[0057] Step 603: Based on the multivariate risk prediction coefficients and the preset risk prediction thresholds, mark the risk warning level, the type of risk event, and its severity;

[0058] Step 604 involves conducting an in-depth analysis of the identified risk events, identifying the underlying risk factors and causes, explaining and elaborating on the prediction results of the risk events, and generating a risk assessment report based on the risk assessment results, providing targeted recommendations and measures.

[0059] Secondly, a data analysis system for automobile insurance business is provided for implementing data analysis methods for automobile insurance business. The system includes a business management center, which is communicatively connected to a data acquisition and preprocessing module, a feature analysis module, a risk assessment model construction module, a diversified assessment model construction module, a risk early warning module, a real-time monitoring module, and a risk assessment report generation module. The modules are electrically connected to each other.

[0060] The data acquisition and preprocessing module is used to collect raw data related to auto insurance business from various data sources and preprocess the collected data to ensure the quality and accuracy of the data, provide a reliable data foundation for subsequent analysis, improve the efficiency and accuracy of data analysis, and reduce analytical bias caused by data problems.

[0061] The feature analysis module is used to extract business risk indicators and claims risk indicators from the original data of auto insurance business, and analyze the sub-indicator features of business risk indicators and claims risk indicators to improve the predictive ability of the model and enhance the model's sensitivity and accuracy in risk identification.

[0062] The risk assessment model construction module is used to assess business risks and claims risks, analyze the historical trends and current status of auto insurance business, and promptly identify potential risk points.

[0063] The diversified assessment model construction module is used to combine business risk indicators and claims risk indicators to build a diversified assessment model, predict and assess risks in auto insurance business, provide comprehensive risk assessment capabilities, cover different types of risk events, improve the accuracy and timeliness of risk assessment, and provide strong support for decision-making.

[0064] The risk warning module is used to combine with the diversified assessment model construction module to divide the risk warning level into low risk warning level, medium risk warning level and high risk warning level according to the risk assessment results, so as to realize the hierarchical management of risk and the rapid response to high risk events.

[0065] The real-time monitoring module is used to establish a real-time monitoring system, receive and process new data streams in real time, monitor and analyze the new data streams in real time, identify potential risk events, realize real-time monitoring and early warning of risks, reduce the impact of risk events on business, improve the response speed of risk response, and enhance the stability and security of business operations.

[0066] The risk assessment report generation module is used to conduct in-depth analysis of identified risk events, extract key risk factors and characteristics, and generate detailed risk assessment reports based on the output results of a diversified assessment model. The report content includes risk factor analysis, interpretation of prediction results, and recommended measures, providing decision-makers with comprehensive risk assessment information to help them understand the nature, scope of impact, and possible consequences of risk events, and to provide targeted recommendations to support decision-makers in formulating effective risk management strategies.

[0067] Thirdly, a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, is used to implement a data analysis method for automobile insurance business.

[0068] Due to the adoption of the above technical solution, the technical progress achieved by this invention compared to the prior art is as follows:

[0069] 1. This invention provides a data analysis method, system, and storage medium for auto insurance business. By constructing a diversified assessment model, it conducts a comprehensive assessment of different types of risk events, greatly enhancing the comprehensiveness and accuracy of risk assessment. It can also more accurately capture risk signals, provide early warnings of potential risks, and offer more reliable decision-making basis. At the same time, it promotes the continuous improvement and optimization of the risk assessment system and drives the intelligent development of auto insurance business.

[0070] 2. This invention provides a data analysis method, system, and storage medium for auto insurance business. By constructing a real-time monitoring system, it enables real-time monitoring and analysis of new data streams, integrates diversified assessment models, instantly identifies potential risk events, and triggers corresponding early warning mechanisms to ensure the timeliness and effectiveness of risk response. The construction of the real-time monitoring system significantly improves the response speed of risk management, helps reduce the impact of risk events on business, and ensures the stable operation of business.

[0071] 3. This invention provides a data analysis method, system, and storage medium for auto insurance business. By automatically generating detailed risk assessment reports, including risk factor analysis, interpretation of prediction results, and recommended measures, it greatly improves decision-making efficiency and quality. It can quickly obtain comprehensive and accurate risk assessment information, understand the nature, scope of impact, and possible consequences of risk events, and through precise analysis of market demand and customer behavior, it can develop insurance products that better meet market needs and improve customer satisfaction. Attached Figure Description

[0072] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0073] Figure 1 This is a flowchart of the method of the present invention;

[0074] Figure 2 This is a flowchart illustrating the process of obtaining the business risk assessment index and the compensation risk assessment index for this invention.

[0075] Figure 3 This is a flowchart illustrating the process of obtaining the meta-risk prediction coefficient of this invention.

[0076] Figure 4 This is a diagram showing the module configuration of the present invention. Detailed Implementation

[0077] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0078] Example 1, such as Figure 1-3 As shown, this invention provides a data analysis method for automobile insurance business, including the following steps:

[0079] Step 1: Obtain and integrate raw auto insurance business data from different channels and sources from the insurance company's internal system, process the data, and integrate it to obtain an insurance business sequence table. Based on the insurance company's internal CRM system, claims system, and financial system, and in conjunction with the bank's credit system, obtain data access permissions and determine the categories of the raw auto insurance business data. The raw auto insurance business data includes customer information, vehicle information, policy details, claims records, claims process, customer service records, and credit records. Among them, customer information includes name, contact information, and address; vehicle information includes license plate number, vehicle model, purchase date, and vehicle value; policy details include insurance type, insured amount, insurance period, and premium; claims records include claims amount, claims time, and claims reason; and the claims process includes reporting time and investigation records. The system extracts the required categories of raw auto insurance business data from various systems using ETL tools. Data backups are performed to prevent data loss or damage. The collected raw auto insurance business data is cleaned, including removing duplicate records, handling missing values, and correcting errors to improve data quality. Data from different sources is converted to a unified format and standard for subsequent processing and analysis. The cleaned and converted data is integrated into a unified data warehouse, and an insurance business sequence table is constructed. The integrated data is then populated according to the sequence table structure, and the populated sequence table is verified by comparing it with the original auto insurance business data to ensure data accuracy and completeness.

[0080] Step 2: Based on the original data of auto insurance business from the insurance business sequence table, identify and analyze business risk indicators and claims risk indicators. Based on the data analysis results, extract sub-indicator features of business risk indicators and claims risk indicators. Business risk indicators are used to analyze the historical trends and current status of auto insurance business, understanding the overall business development and risks. Claims risk indicators are used for monitoring and early warning, promptly identifying potential risk points. Review the insurance business sequence table to clarify its data structure and content, define business risks, and identify business risk indicators based on the data in the insurance business sequence table. These business risk indicators include business growth rate, premium growth rate, customer retention rate, renewal rate, and insurance penetration. The data is analyzed to determine the business risk indicators and their historical trends. Further refinement of these indicators is achieved by extracting their sub-indicator characteristics. The sub-indicators of business growth rate include annual policy quantity growth rate and annual premium income growth rate. Sub-indicators of premium growth rate include new product sales growth rate and supplementary protection sales growth rate. Sub-indicators of customer retention rate include annual customer retention rate and high-value customer retention rate. Sub-indicators of renewal rate include renewal rates for different product lines and the average number of policies per renewing customer. Sub-indicators of insurance penetration rate include regional insurance coverage rate and insurance participation rate. The annual policy quantity growth rate is used to compare the growth rate of new policies signed each year. The annual premium income growth rate is used to assess the average number of policies signed each year. Annual premium income growth; new product sales growth rate is used to assess the market acceptance and growth potential of newly launched insurance products; supplementary protection sales growth rate is used to analyze the sales growth of additional supplementary protection products and measure the market competitiveness of the product portfolio; annual customer retention rate is used to analyze the proportion of customers who continue to purchase policies each year; high-value customer retention rate is used to analyze the retention rate of high-value customers who make significant contributions; renewal rate of different product lines is used to compare the renewal situation of different insurance product lines; average number of policies per renewing customer is used to analyze the average number of policies held by renewing customers and assess customers' diversified needs and insurance utilization; regional insurance coverage rate is used to measure the insurance product usage rate in a selected region; insurance participation rate is used to analyze... Based on the proportion of customers participating in insurance, potential market growth opportunities are identified, claims risks are defined, and claims risk indicators are identified based on data from the insurance business sequence table. These indicators include the loss ratio, claims processing cycle, and average claim amount per case. A claims risk monitoring and early warning mechanism is established, and the claims risk indicators are further refined, their sub-indicator characteristics are extracted, and potential risk points of the claims risk indicators are analyzed in depth. The sub-indicator characteristics of the loss ratio are the basic loss ratio and the loss ratio of different product lines; the sub-indicator characteristics of the claims processing cycle are the average claims processing time and the processing time of different types of claims; and the sub-indicator characteristics of the average claim amount per case are the average claim amount per insurance type and the maximum difference in claim amount per case per insurance type. The basic loss ratio is used to calculate the ratio of total claims amount to total premium income.The loss ratios of different product lines are used to compare the loss ratios of different insurance products and identify high-risk product lines; the average claims processing time is used to analyze the average time from claim application to payment completion and to assess claims processing efficiency; the processing times of different types of claims are used to compare the average processing times of different types of claims and to identify potential processing bottlenecks and areas for improvement; the average payout per case is used to compare the average payout amount of different types of claims and to identify the characteristics of high-payout cases; the maximum difference in payout per case is used to analyze the distribution of payout amounts and to identify potentially high-risk cases.

[0081] Step 3: Combining the original data of auto insurance business, conduct multi-dimensional analysis of business risk indicators and claims risk indicators, and construct business risk prediction models and claims risk prediction models respectively. Obtain business risk assessment indices and claims risk assessment indices. Extract feature data of business risk indicators and claims risk indicators from the original data of auto insurance business in the insurance business sequence table to obtain business risk feature datasets and claims risk feature datasets. Divide the business risk feature datasets and claims risk feature datasets respectively to obtain training sets and test sets. Based on the business risk indicators, extract their sub-indicator features. Use the training set data of the business risk feature dataset to train the model, construct the business risk prediction model, and use the business risk features... The test set of the dataset is used to evaluate and test the business risk prediction model to analyze the risk status of the auto insurance business. Based on the claims risk index, its sub-indicator features are extracted. The model is trained using the training set data of the claims risk feature dataset to build a claims risk prediction model. The test set of the claims risk feature dataset is used to evaluate and test the claims risk prediction model to predict claims risk. The business risk prediction model and the claims risk prediction model are analyzed to obtain the business risk assessment index and the claims risk assessment index. The business risk assessment index is used to quantify the level of business risk, and the claims risk assessment index is used to quantify the level of claims risk. Risk assessment and monitoring are carried out based on the results of the business risk assessment index and the claims risk assessment index.

[0082] Furthermore, the formula for calculating the business risk assessment index is as follows:

[0083]

[0084] Among them, BI is the Business Risk Assessment Index, and G... i It is the business growth rate at the i-th time point. σ(R) is the average business growth rate, R is the customer retention rate, σ(R) is the standard deviation of the customer retention rate, max(R) is the maximum value of the customer retention rate, min(R) is the minimum value of the customer retention rate, PR is the insurance penetration rate, which measures the degree of popularity of insurance products in a specific market, n is the number of time points, and μ(R) is the average customer retention rate. It should be noted that the smaller the BI, the higher the business risk; the larger the BI, the lower the business risk.

[0085] The formula for calculating the payout risk assessment index is as follows:

[0086]

[0087] max(V)=max(V1,V2,…,V K );

[0088] CI stands for Claims Risk Assessment Index. j P is the payout amount for the j-th claim. j Is with C j The corresponding premium income, m is the number of claims, used as the exponent in the square root, CT is the claims processing time, which is the average time from claim application to payment completion, max(CT) is the maximum claim processing time, and min(CT) is the minimum claim processing time. V represents the average payout per case, and max(V) represents the maximum variance of the average payout per case, indicating the volatility of the payout. k It is the variance of the average claim amount, K is the number of time periods, and o k It refers to the number of cases within a given time period, C. jk It is the payout amount for the j-th claim in the k-th time period, μ(C k CI represents the average payout amount over a given period. It should be noted that a more negative CI indicates a higher payout risk, while a more positive CI indicates a lower payout risk.

[0089] Step 4: Based on business characteristics and risk assessment needs, comprehensively analyze business risk indicators and claims risk indicators to establish a diversified assessment model. Combine the business risk assessment index and claims risk assessment index to obtain diversified risk prediction coefficients. Comprehensively assess the impact of different risk factors on auto insurance business. Extract feature datasets of business risk and claims risk from the original data of auto insurance business, and divide the extracted feature datasets into training set and test set. Using the training set data and the features of business risk indicators and claims risk indicators, use a multiple linear regression model to establish a diversified assessment model. Use the test set data to test and analyze the diversified assessment model. Comprehensively analyze the diversified assessment model, and combine the business risk assessment index and claims risk assessment index to obtain diversified risk prediction coefficients. Use the diversified risk prediction coefficients and diversified assessment model to conduct risk assessment on the current business and analyze the comprehensive impact of different risk factors on auto insurance business.

[0090] Furthermore, the formula for calculating the multivariate risk prediction coefficient is as follows:

[0091]

[0092] Where MRC is the multivariate risk prediction coefficient, BI l It is the l-th business risk assessment index, CI l It is the l-th compensation risk assessment index, and v is the total number of features of business risk indicators and compensation risk indicators. It is the average business growth rate, used in the limit function to represent the long-term growth trend. It should be noted that a higher MRC value indicates that business risk and payout risk are well balanced and risk management is effective, while a lower MRC value indicates an unbalanced risk situation that requires further risk analysis and management.

[0093] Step 5: Based on the multivariate risk prediction coefficients and the original data of the auto insurance business, classify the risk warning levels according to the recorded risk events, and set corresponding risk prediction thresholds for the risk warning levels. Based on the collected original data of the auto insurance business, analyze the recorded risk events, determine the types and severity of the risk events. The types of risk events include claims events, fraudulent behavior, and high-risk driving behavior. Based on the determined types of risk events, combined with the distribution of the multivariate risk prediction coefficients and the severity of the risk events, classify the risk warning levels as low-risk warning level, medium-risk warning level, and high-risk warning level. Based on the risk prediction values ​​of the multivariate risk prediction coefficients, set corresponding risk prediction thresholds for each risk warning level.

[0094] Multiple risk warning levels correspond to multiple risk prediction thresholds, which include upper and lower thresholds.

[0095] Multiple risk warning levels and multiple risk prediction thresholds satisfy the following relationship:

[0096] Low risk warning level MRC≥MRC zy A lower risk prediction value indicates that the policyholder has a lower risk profile and is less likely to experience claims, fraud, or high-risk driving behavior.

[0097] Medium risk warning level MRC gy ≤MRC <MRC zy The risk prediction value is at a moderate level, indicating that the policyholder faces certain risks and their claims and driving behavior need to be monitored.

[0098] High-risk warning level 0 <MRC<MRC gy A high risk prediction value indicates that the policyholder faces a higher risk of claims, fraud, or high-risk driving behavior.

[0099] Where MRC represents the multivariate risk prediction coefficient, MRC zy MRC sets the lower threshold for low-risk warning levels and the upper threshold for medium-risk warning levels. gy These are the lower threshold corresponding to the medium-risk warning level and the upper threshold corresponding to the high-risk warning level;

[0100] Step 6: Based on the established diversified assessment model, construct a real-time monitoring system to monitor new data flows and identify risk events, generating risk assessment reports, including risk factor analysis, interpretation of prediction results, and recommended measures to optimize execution strategies. Establish a data pipeline to collect real-time data flows from the auto insurance business, including policy information, claim records, driving behavior data, etc. Ensuring the real-time nature and completeness of data is the foundation of system construction. Integrate the diversified assessment model into the monitoring system, use the diversified assessment model to perform real-time analysis of the auto insurance business data flows, identify potential risk events, and mark the risk warning level, risk event type, and severity according to the diversified risk prediction coefficients and preset risk prediction thresholds. Conduct in-depth analysis of the identified risk events to determine the underlying risk factors and causes, interpret and explain the prediction results of risk events, and based on the risk assessment results, generate a risk assessment report and provide targeted recommended measures.

[0101] Example 2, as Figure 4 As shown, based on Embodiment 1, the present invention also provides a data analysis system for automobile insurance business, which is used to implement a data analysis method for automobile insurance business. The system includes a business management center, which is communicatively connected to a data acquisition and preprocessing module, a feature analysis module, a risk assessment model construction module, a diversified assessment model construction module, a risk warning module, a real-time monitoring module, and a risk assessment report generation module. The modules are electrically connected to each other.

[0102] The data acquisition and preprocessing module is used to collect raw data related to auto insurance business from various data sources and preprocess the collected data to ensure data quality and accuracy, provide a reliable data foundation for subsequent analysis, improve the efficiency and accuracy of data analysis, and reduce analytical bias caused by data problems.

[0103] The feature analysis module is used to extract business risk indicators and claims risk indicators from the raw data of auto insurance business, and analyze the sub-indicator characteristics of business risk indicators and claims risk indicators to improve the predictive ability of the model and enhance the model's sensitivity and accuracy in risk identification.

[0104] The risk assessment model building module is used to assess business risks and claims risks, analyze the historical trends and current status of the auto insurance business, and identify potential risk points in a timely manner.

[0105] The diversified assessment model building module is used to combine business risk indicators and claims risk indicators to build diversified assessment models, predict and assess risks in auto insurance business, provide comprehensive risk assessment capabilities, cover different types of risk events, improve the accuracy and timeliness of risk assessment, and provide strong support for decision-making.

[0106] The risk warning module is used to combine a diversified assessment model with the module to classify low-risk, medium-risk, and high-risk warning levels based on the risk assessment results, thereby achieving hierarchical risk management and rapid response to high-risk events.

[0107] The real-time monitoring module is used to establish a real-time monitoring system, receive and process new data streams in real time, monitor and analyze new data streams in real time, identify potential risk events, realize real-time monitoring and early warning of risks, reduce the impact of risk events on business, improve the response speed of risk response, and enhance the stability and security of business operations.

[0108] The risk assessment report generation module is used to conduct in-depth analysis of identified risk events, extract key risk factors and characteristics, and generate detailed risk assessment reports based on the output of a diversified assessment model. The report content includes risk factor analysis, interpretation of prediction results, and recommended measures, providing decision-makers with comprehensive risk assessment information to help them understand the nature, scope of impact, and possible consequences of risk events, and to provide targeted recommendations to support decision-makers in formulating effective risk management strategies.

[0109] Example 3: Based on Examples 1-2, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method mentioned above.

[0110] Generally, the computer instructions for implementing the method of the present invention can be carried on any combination of one or more computer-readable storage media. Non-transitory computer-readable storage media can include any computer-readable medium except for signals that are temporarily propagating themselves.

[0111] Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

[0112] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—as well as conventional procedural programming languages—such as the "C" language or similar programming languages. In particular, Python, suitable for neural network computation, and platform frameworks based on TensorFlow, PyTorch, etc., can be used. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or connected to an external computer (e.g., via the Internet using an Internet service provider).

[0113] For details regarding the aforementioned non-transitory computer-readable storage medium, please refer to the implementation details and beneficial effects of the methods mentioned above, which will not be repeated here.

[0114] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A data analysis method for automobile insurance business, characterized by: The following steps are involved: Step 1: Obtain and integrate original automobile insurance business data from different channels and sources from the insurance company's internal system, process the data, and integrate to obtain an insurance business sequence table; Step 2, based on the original data of automobile insurance business in the insurance business sequence table, identify and analyze the business risk indicators and the compensation risk indicators, and extract the sub-indicator characteristics of the business risk indicators and the compensation risk indicators based on the data analysis results; Step 3: Combine the original data of automobile insurance business, conduct multi-dimensional analysis on business risk indicators and claim risk indicators, respectively build business risk prediction model and claim risk prediction model, and obtain business risk assessment index and claim risk assessment index. In step 3, the process of obtaining business risk assessment index and claim risk assessment index is as follows: Step 301, extracting characteristic data of business risk indicators and claim risk indicators from the original data of automobile insurance business in the insurance business sequence table, obtaining a business risk characteristic data set and a claim risk characteristic data set, and dividing the business risk characteristic data set and the claim risk characteristic data set respectively to obtain a training set and a test set; Step 302: based on the business risk indicator, extract its sub-indicator features, use the training set data of the business risk feature data set to perform model training, build a business risk prediction model, and use the test set of the business risk feature data set to evaluate and test the business risk prediction model to analyze the risk status of the automobile insurance business; Step 303: based on the claim risk indicator, extract its sub-indicator features, use the training set data of the claim risk feature data set to perform model training, build a claim risk prediction model, and use the test set of the claim risk feature data set to evaluate the claim risk prediction model to predict the claim risk; Step 304: Analyze the business risk prediction model and the compensation risk prediction model to obtain a business risk assessment index and a compensation risk assessment index; Step 305, performing risk assessment and monitoring based on the results of the business risk assessment index and the claim risk assessment index; The calculation formula of the business risk assessment index is: Among them, BI is the business risk assessment index, G i is the business growth rate at the i-th time point, is the average value of the business growth rate, R is the customer retention rate, σ(R) is the standard deviation of the customer retention rate, max(R) is the maximum value of the customer retention rate, min(R) is the minimum value of the customer retention rate, PR is the insurance penetration rate, n is the number of time points, and μ(R) is the average value of the customer retention rate; The calculation formula of the compensation risk assessment index is: max(V)=max(V1,V2,…,V K ); Among them, CI is the compensation risk assessment index, C j is the payment amount of the jth claim case, P j is with C j The corresponding premium income, m is the number of claims, used as the square root index, CT is the claim cycle time, that is, the average time from claim application to payment completion, max(CT) is the maximum value of the claim cycle time, min(CT) is the minimum value of the claim cycle time, is the average compensation amount per case, max(V) is the maximum variance of the average compensation amount per case, indicating the volatility of the compensation amount, V k is the variance of the average compensation amount per case, K is the number of time periods, o k is the number of cases in the time period, C jk is the amount of compensation paid for the jth claim in the kth time period, μ(C k ) is the average amount of claims paid during the time period; Step 4: Based on the business characteristics and risk assessment requirements, comprehensively analyze the business risk indicators and the compensation risk indicators, establish a diversified assessment model, and combine the business risk assessment index and the compensation risk assessment index to obtain a multivariate risk prediction coefficient, and comprehensively assess the impact of different risk factors on the automobile insurance business. In step 4, the process of obtaining the multivariate risk prediction coefficient is as follows: Step 401, extracting a feature data set of business risk and claim risk from the original data of automobile insurance business, and dividing the extracted feature data set into a training set and a test set; Step 402, using the training set data combined with the characteristics of the business risk indicators and the compensation risk indicators, using a multivariate linear regression model to establish a diversified evaluation model, and using the test set data to test and analyze the diversified evaluation model; Step 403, comprehensively analyze the diversified assessment model, and combine the business risk assessment index and the claim risk assessment index to obtain a multivariate risk prediction coefficient. Use the multivariate risk prediction coefficient and the diversified assessment model to conduct risk assessment on the current business and analyze the comprehensive impact of different risk factors on the automobile insurance business. The calculation formula of the multivariate risk prediction coefficient is: Among them, MRC is the multivariate risk prediction coefficient, BI l is the lth business risk assessment index, CI l is the lth compensation risk assessment index, v is the total number of characteristics of business risk indicators and compensation risk indicators, It is the average value of the business growth rate and is used in the limit function to indicate the long-term growth trend; Step 5, based on the multivariate risk prediction coefficient and the original data of the automobile insurance business, the risk warning level is divided according to the recorded risk events, and the corresponding risk prediction threshold is set for the risk warning level; Step 6: Build a real-time monitoring system based on the established diversified assessment model to monitor new data streams and identify risk events, and generate risk assessment reports to optimize execution strategies.

2. The data analysis method for automobile insurance business according to claim 1, characterized in that: In step 1, the process of obtaining the insurance business sequence table is as follows: Step 101, based on the insurance company's internal CRM system, claims system and financial system, and in combination with the bank's credit system, obtain data access rights and determine the category of the original data of the automobile insurance business, which includes customer information, vehicle information, policy details, payment records, claims process, customer service records and credit records; Step 102, using ETL tools to extract the required category of automobile insurance business original data from each system, and back up the data; Step 103, clean the collected original data of automobile insurance business, convert the data from different sources into a unified format and standard, and integrate the cleaned and converted data into a unified data warehouse; Step 104, constructing an insurance business sequence table, filling the integrated data according to the structure of the sequence table, and verifying the filled sequence table by comparing the original data of the automobile insurance business.

3. The data analysis method for automobile insurance business according to claim 2, characterized in that: In step 2, the process of extracting the characteristics of the business risk index and the sub-indicator of the claim risk index is as follows: Step 201, review the insurance business sequence table to clarify the data structure and content of the insurance business sequence table; Step 202, define business risks and identify business risk indicators based on the data in the insurance business sequence table, where the business risk indicators include business growth rate, premium growth rate, customer retention rate, renewal rate and insurance penetration rate, and analyze the historical trends of the business risk indicators, further refine the business risk indicators, extract their sub-indicator features, and deeply analyze the change trends of the business risk indicators; Step 203, the sub-indicator characteristics of the business growth rate are the annual policy quantity growth rate and the annual premium income growth rate, the sub-indicator characteristics of the premium growth rate are the new product sales growth rate and the additional protection sales growth rate, the sub-indicator characteristics of the customer retention rate are the annual customer retention rate and the high-value customer retention rate, the sub-indicator characteristics of the renewal rate are the renewal rates of different product lines and the average number of policies of the renewal customers, and the sub-indicator characteristics of the insurance penetration rate are the regional insurance coverage rate and the insurance participation rate; Step 204, define the claim risk and identify the claim risk indicators based on the data in the insurance business sequence table, where the claim risk indicators include the claim ratio, claim settlement cycle, and average claim amount per case, and establish a claim risk monitoring and early warning mechanism, further refine the claim risk indicators, extract the characteristics of its sub-indicators, and deeply analyze the potential risk points of the claim risk indicators; Step 205, the sub-indicator characteristics of the claims ratio are the basic claims ratio and the claims ratios of different product lines, the sub-indicator characteristics of the claims cycle are the average claims processing time and the processing time of different types of claims, and the sub-indicator characteristics of the average compensation amount per case are the average compensation amount per insurance type and the maximum difference between the insurance type and a single case compensation.

4. The data analysis method for automobile insurance business according to claim 1, characterized in that: In step 5, the risk prediction threshold setting process of the risk warning level is as follows: Step 501, based on the collected original data of automobile insurance business, analyze the recorded risk events to determine the type and severity of the risk events, where the types of risk events include compensation events, fraudulent behavior, and high-risk driving behavior; Step 502, according to the determined type of risk event, combined with the distribution of multivariate risk prediction coefficients and the severity of the risk event, the risk warning level is divided into a low risk warning level, a medium risk warning level and a high risk warning level; Step 503, based on the risk prediction value of the multivariate risk prediction coefficient, set a corresponding risk prediction threshold for each risk warning level; A plurality of the risk warning levels correspond to a plurality of the risk prediction thresholds, wherein the risk prediction thresholds include an upper threshold and a lower threshold; The multiple risk warning levels and the multiple risk prediction thresholds satisfy the following relationship: Low risk warning level MRC ≥ MRC zy ; Medium risk warning level MRC gy ≤MRC <MRC zy ; High risk warning level 0 <MRC<MRC gy ; Among them, MRC represents the multivariate risk prediction coefficient, MRC zy The lower threshold corresponding to the low risk warning level and the upper threshold corresponding to the medium risk warning level, MRC gy It is the lower threshold corresponding to the medium risk warning level and the upper threshold corresponding to the high risk warning level.

5. A data analysis system for automobile insurance business, used to implement a data analysis method for automobile insurance business as claimed in any one of claims 1 to 4, comprising a business management center, characterized in that: The business management center is communicatively connected with a data acquisition and preprocessing module, a feature analysis module, a risk assessment model building module, a diversified assessment model building module, a risk warning module, a real-time monitoring module, and a risk assessment report generation module, wherein the modules are electrically connected with each other; The data collection and preprocessing module is used to collect raw data related to automobile insurance business from various data sources and preprocess the collected data; The feature analysis module is used to extract business risk indicators and claim risk indicators from the original data of automobile insurance business, and analyze the sub-indicator characteristics of the business risk indicators and claim risk indicators; The risk assessment model building module is used to assess business risks and claims risks, analyze the historical trends and current status of automobile insurance business, and promptly identify potential risk points; The diversified assessment model building module is used to combine business risk indicators and claim risk indicators to build a diversified assessment model to predict and assess risks in automobile insurance business; The risk warning module is used to combine the diversified assessment model construction module, divide the risk warning level into low risk warning level, medium risk warning level and high risk warning level according to the risk assessment results, realize the hierarchical management of risks, and quickly respond to high risk events; The real-time monitoring module is used to establish a real-time monitoring system, receive and process new data streams in real time, monitor and analyze new data streams in real time, identify potential risk events, and achieve real-time monitoring and early warning of risks; The risk assessment report generation module is used to conduct in-depth analysis of identified risk events, extract key risk factors and features, and generate a detailed risk assessment report based on the output results of the diversified assessment model.

6. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it is used to implement a data analysis method for automobile insurance business as described in any one of claims 1 to 4.

Images