CN118037460A

CN118037460A - Method and system for evaluating insurance product

Info

Publication number: CN118037460A
Application number: CN202410312480.9A
Authority: CN
Inventors: 肖建; 陈镁琦
Original assignee: Xinjiang Yisheng Xinchuangzhan Technology Co ltd
Current assignee: Xinjiang Yisheng Xinchuangzhan Technology Co ltd
Priority date: 2024-03-19
Filing date: 2024-03-19
Publication date: 2024-05-14

Abstract

The invention relates to the technical field of data mining, in particular to an evaluation method and system of an insurance product. The method comprises the following steps: acquiring historical data of insurance products, carrying out claim settlement risk evaluation, and acquiring historical claim settlement risk data; constructing a claim risk evaluation model based on the historical claim risk data; acquiring insurance product data to be evaluated, performing claim risk evaluation, and acquiring claim risk evaluation data; performing coverage evaluation on the historical data of the insurance product to obtain coverage evaluation data; calculating product default interaction frequency of the insurance product data to be evaluated, obtaining product default interaction frequency data to be evaluated, and performing customer interaction experience evaluation to obtain interaction experience evaluation data; and carrying out comprehensive evaluation on the insurance product according to the claim risk evaluation data, the coverage area evaluation data and the interactive experience evaluation data, thereby obtaining an insurance product evaluation report. The invention can improve the accuracy of the insurance product evaluation result.

Description

Method and system for evaluating insurance product

Technical Field

The invention relates to the technical field of data mining, in particular to an evaluation method and system of an insurance product.

Background

Conventional insurance product assessment methods are typically based on historical data and statistical models that make it difficult to fully understand the personalized needs and risk characteristics of the customer. Since these models are typically based on past data and statistical analysis, future risk situations may not be accurately predicted, especially with limited ability to evaluate new risks or extreme events. In addition, existing insurance product assessment systems often focus on only some traditional risk factors, such as gender, occupation, etc., but ignore many other factors that may affect the customer experience, such as customer communication frequency, product payment period, etc. This results in a lack of comprehensiveness and accuracy of the assessment results, which do not adequately reflect the personalized risk of the customer.

Disclosure of Invention

Accordingly, the present invention is directed to a method and system for evaluating an insurance product, so as to solve at least one of the above-mentioned problems.

In order to achieve the above object, a method for evaluating an insurance product includes the steps of:

Step S1: acquiring historical data of the insurance product, and carrying out claim settlement risk evaluation on the historical data of the insurance product so as to acquire historical claim settlement risk data; constructing a claim risk evaluation model based on the historical claim risk data;

Step S2: acquiring insurance product data to be evaluated, and performing claim risk evaluation on the insurance product data to be evaluated through a claim risk evaluation model, so as to acquire claim risk evaluation data;

step S3: extracting customer characteristics of the historical data of the insurance product so as to obtain historical customer data, and classifying customer loyalty of the historical customer data so as to obtain customer loyalty data; performing coverage evaluation on insurance product data to be evaluated according to the customer loyalty data, thereby obtaining coverage evaluation data;

Step S4: calculating product default interaction frequency of the insurance product data to be evaluated so as to obtain product default interaction frequency data to be evaluated, and performing customer interaction experience evaluation of the product default interaction frequency data to be evaluated and insurance product historical data so as to obtain interaction experience evaluation data;

Step S5: and carrying out comprehensive evaluation on the insurance product according to the claim risk evaluation data, the coverage area evaluation data and the interactive experience evaluation data, thereby obtaining an insurance product evaluation report.

The invention builds a claim risk evaluation model by analyzing the historical claim data of the insurance product. Such models can identify and quantify different types of risk, providing an important basis for subsequent insurance product assessment. By knowing the historical claims data, future potential claims risk can be better predicted. By conducting claim risk assessment on the insurance product data to be assessed, the risk level of the product can be quantified. This helps the insurance company better understand the risk characteristics of each product, thereby making pricing and claim policies more efficient. By performing customer feature extraction and loyalty classification on historical customer data, customer behavior and preferences may be better understood. Such evaluation may help insurance companies to better understand their existing customer groups and design more attractive insurance products and services to improve customer loyalty. By calculating the default interaction frequency and customer interaction experience evaluation of the product, the interactivity and customer friendliness of the insurance product can be evaluated. This helps insurance companies design more attractive and easy to use products, enhancing customer experience and thus customer satisfaction and loyalty. Comprehensively considering the risk evaluation, coverage evaluation and interactive experience evaluation data of the claims, and obtaining a comprehensive evaluation report of the insurance product. Such comprehensive assessment can help insurance companies to fully understand the advantages and improvement spaces of products, and provide important references for business decisions.

Optionally, step S1 specifically includes:

step S11: acquiring historical data of insurance products;

step S12: extracting deadline characteristics of historical data of the insurance products so as to obtain historical long-term insurance product data and historical short-term insurance product data;

Step S13: performing long-term product claim settlement risk evaluation on the historical long-term insurance product data so as to obtain long-term product claim settlement risk data;

Step S14: performing short-term product claim settlement risk evaluation on the historical short-term insurance product data so as to obtain short-term product claim settlement risk data;

step S15: combining the long-term product claim risk data and the short-term product claim risk data, thereby obtaining historical claim risk data;

Step S16: and constructing a claim risk evaluation model based on the historical claim risk data.

By acquiring the historical data of the insurance product, the invention can establish a comprehensive data set which comprises the past insurance product information, insurance policy details, claim records and the like. This provides the basis data for the subsequent steps so that the model can take into account the influence of the historical information in the risk assessment of claims. Through deadline feature extraction, insurance product data can be divided into a long term and a short term, and risks of the two products can be analyzed in a targeted manner. This helps to more accurately assess the difference in terms of claim risk for long-term and short-term insurance products. By conducting claim risk assessment on historical long-term insurance product data, the potential risk of long-term products can be identified and quantified. This helps the insurance company better understand and manage the risk of long-term products, taking corresponding insurance policies. Similar to long-term products, the risk assessment of claims to historical short-term insurance product data helps identify and quantify the risk of short-term products. This allows the insurer to better understand and address the claims risk that short term products may face. By combining the claims risk data for both long-term and short-term products, a more comprehensive historical claims risk data set can be created. The risk factors of long-term and short-term products are comprehensively considered, and more abundant information is provided for constructing a comprehensive claim risk evaluation model. And constructing a claim risk assessment model, wherein the model can predict the claim risk of future insurance products based on the historical claim risk data. This provides a powerful tool for insurance companies to price products more accurately, formulate risk management policies, and improve overall business efficiency.

Optionally, step S13 specifically includes:

step S131: extracting features of the historical long-term insurance product data so as to obtain historical long-term claim settlement data and historical long-term premium payment data;

step S132: calculating the same product claim settlement probability according to the historical long-term claim settlement data so as to obtain product claim settlement probability data, and carrying out statistical analysis on the product claim settlement probability data so as to obtain a product claim settlement risk threshold value;

step S133: classifying the long-term claim settlement data according to the product claim settlement risk threshold value, so as to obtain high-claim settlement risk long-term product data and low-claim settlement risk long-term product data;

Step S134: counting the number of missed customer payments according to the historical long-term premium payment data, thereby obtaining high-frequency missed customer payment data and low-frequency missed customer payment data;

Step S135: according to the high-frequency customer missing payment data and the low-frequency customer missing payment data, classifying the customer risk of the historical long-term insurance product data, so as to obtain high-customer-risk long-term product data and low-customer-risk long-term product data;

Step S136: intersection operation is carried out on the long-term product data with high claim risk and the long-term product data with high customer risk, thereby obtaining long-term product claim high risk data; performing intersection operation on the low-risk long-term product data and the low-client-risk long-term product data, so as to obtain low-risk data of the long-term product claims;

step S137: and merging the low risk data of the long-term product claim and the high risk data of the long-term product claim, thereby obtaining the risk data of the long-term product claim.

By means of feature extraction, the method and the device can extract useful features from historical data for subsequent risk assessment and classification. The calculation of the claim probability and the risk threshold can help the insurance company to better know the risk level of the product and provide basis for risk classification. Classifying the products according to the risk level can help insurance companies to more pertinently formulate a risk management strategy and improve the effect of risk control. The statistics of the missing payment times of the clients can help the insurance company to know the payment condition of the clients and identify the high-risk clients, so that corresponding measures are taken to reduce risks. Classifying clients according to risk levels can help insurance companies to better understand risk characteristics of client groups, and client management and risk control can be performed pertinently. By performing intersection operation on the claim risk and the customer risk, the high-risk product and the customer can be identified more accurately, and finer data support is provided for risk management. The low-risk data and the high-risk data are combined to obtain a complete risk data set, so that a more comprehensive risk assessment basis is provided for insurance companies, and decisions such as insurance product design and pricing are supported.

Optionally, step S14 specifically includes:

Step S141: extracting features of the historical short-term insurance product data so as to obtain historical short-term claim settlement data and customer age data;

Step S142: counting the frequency of the short-term claims to obtain high-frequency short-term claims and low-frequency short-term claims;

step S143: classifying and calculating the customer age data so as to obtain senior customer age data and senior customer age data;

Step S144: according to the age data of the senior clients and the age data of the senior clients, extracting the age-limited insurance products from the historical short-term insurance product data, so as to obtain short-term age-limited insurance product data; performing intersection operation according to the low-frequency short-term claim settlement data and the short-term age limit insurance product data, so as to obtain short-term product claim settlement low risk data;

Step S145: carrying out customer association on the high-frequency short-term claim settlement data and the advanced customer age data so as to obtain first customer association data, and carrying out high-risk advanced customer product data extraction on the historical short-term insurance product data according to the first customer association data so as to obtain advanced short-term product claim settlement high-risk data;

Step S146: carrying out customer association on the high-frequency short-term claim settlement data and the age data of the low-age customers so as to obtain second customer association data, and extracting high-risk low-age customer product data of the historical short-term insurance product data according to the second customer association data so as to obtain high-risk data of the low-age short-term product claim settlement;

Step S147: and merging the low risk data of the short-term product claims, the high risk data of the advanced short-term product claims and the high risk data of the low-age short-term product claims, thereby obtaining the short-term product claims.

The invention can know the claim settlement condition and age distribution of the clients through the feature extraction, and provides a data basis for subsequent risk analysis and client association. The frequency statistics of claims may help identify products that are frequently claimed, and these data may be considered as potentially high risk factors. Through classification of age data, risk analysis and customized insurance product design can be performed on clients of different age groups. By limiting the age range and the claim settlement frequency, safer insurance products and clients can be extracted, and the risk of claim settlement is reduced. Through association analysis, common characteristics of high-risk client groups can be found, and risk management and product design can be performed in a targeted manner. And combining the data with different risk levels to obtain comprehensive claim risk data, thereby providing support for an insurance company to formulate a more effective risk management strategy.

Optionally, step S3 specifically includes:

Step S31: extracting client characteristics of the insurance product historical data so as to obtain historical client data;

step S32: extracting the participating product data of the historical customer data so as to obtain the participating product data of the customer;

Step S33: counting the participation and protection period of the client participation and protection product data so as to obtain long-term client participation and protection product data and short-term client participation and protection product data, and calculating the participation and protection product proportion of the long-term client participation and protection product data and the short-term client participation and protection product data so as to obtain client participation and protection product proportion data;

Step S34: counting the number of the participating products of the client participating data, thereby obtaining the number of the client participating products;

step S35: performing customer loyalty division on the proportion data of the customer participating products and the number of the customer participating products so as to obtain customer loyalty data;

Step S36: and performing coverage evaluation on the insurance product data to be evaluated according to the customer loyalty data, thereby obtaining coverage evaluation data.

By extracting the client characteristics, the invention can know the personal and insurance related information of the history client and provide a data basis for subsequent analysis. By extracting the data of the customer participating products, the insurance purchasing behavior of the customer can be known, and a customer-product association is formed. Knowing the participation of long-term and short-term customers, calculating the proportion of the participating products helps to know the insurance purchasing preference of the customers, and provides guidance for product design and sales strategies. Through counting the number of the participating products, the insurance combination condition of the clients can be known, and references are provided for product recommendation and pricing. Dividing customers into different loyalty tiers may help companies identify high and low loyalty customers, thereby developing corresponding customer relationship management policies. By evaluating the coverage range of the insurance product to be evaluated, the matching degree of the product and different loyalty clients can be known, and data support is provided for product improvement and positioning.

Optionally, the customer loyalty division in step S35 is specifically:

Carrying out product proportion statistical analysis on the product proportion data of the customer participating in the insurance so as to obtain short-term product high-duty-ratio customer data, long-term product high-duty-ratio customer data, monotone product customer data and same-duty-ratio customer data;

carrying out product quantity statistical analysis on the number of the customer participating and protecting products so as to obtain high-amount participating and protecting number customer data and low-amount participating and protecting number customer data;

performing customer intersection operation according to the long-term product high-duty customer data and the high-amount participating number customer data, thereby obtaining first high-loyalty customer data;

Performing customer intersection operation according to the short-term product high-duty customer data and the high-amount participating number customer data, thereby obtaining second high-loyalty customer data;

Performing customer intersection operation according to the same-duty customer data and the high-amount participating number customer data, thereby obtaining third-highest loyalty customer data;

performing customer intersection operation according to the monotone product customer data and the high-amount participating number customer data, thereby obtaining fourth high-loyalty customer data;

performing customer intersection operation according to the monotone product customer data and the low-amount participating number customer data, thereby obtaining low-loyalty customer data;

And carrying out data combination on the first high-loyalty client data, the second high-loyalty client data, the third high-loyalty client data and the fourth high-loyalty client data so as to obtain high-loyalty client data, and carrying out data combination on the high-loyalty client data and the low-loyalty client data so as to obtain client loyalty data.

These customers in the short-term product high-duty customer data of the present invention may prefer short-term products, may require more frequent service or may be more susceptible to market changes. These customers in the long-term product high-occupancy customer data may prefer long-term products, may require more stable investment, or may pay more attention to long-term security. These customers in monotonic product customer data may have a higher preference for long-term or short-term types of products and may have a more explicit need to focus on their product satisfaction and demand variations. These customers in the same duty cycle customer data may have similar duty cycles across different product types, and may have certain needs and preferences for different types of products. These customers in the high-participation-amount customer data may purchase a variety of products, which may be high-equity customers or customers interested in a variety of guarantees. These customers in the low participating amount customer data may purchase less product, may be new customers, small investors, or customers with poor knowledge of the product. By performing intersection operations on customer data under different conditions, such as long term product high duty cycle intersections with high participating number of customers, a population of customers meeting a plurality of conditions may be found, which customers may have higher loyalty and higher investment requirements. By identifying the intersection of a monotonous product customer and a low participating number of customers, a population of customers may be found that may require additional attention and guidance, who may be less aware of the product or more sensitive to market fluctuations. Combining customer data of different loyalty levels can form one comprehensive customer loyalty data, which is helpful for formulating marketing strategies and service schemes for customers of different loyalty levels.

Optionally, step S36 specifically includes:

Step S361: extracting features of the insurance product data to be evaluated so as to obtain the deadline data of the product to be evaluated and the premium data of the product to be evaluated;

Step S362: extracting the customer history participating product data from the insurance product history data according to the customer loyalty data, thereby obtaining the customer history participating product data;

step S363: calculating the product period duty ratio of the customer history reference product data to obtain customer history product period duty ratio data, and calculating the period coverage of the customer history product period duty ratio data according to the product period data to be evaluated to obtain customer period coverage data;

Step S364: calculating the product premium ratio of the historical participating product data of the client so as to obtain the historical product premium ratio data of the client, and calculating the average premium of the historical participating product data of the client so as to obtain the average premium data of the client;

Step S365: calculating premium dispersion of the premium data of the product to be evaluated and the average premium data of the clients, so as to obtain premium dispersion data; calculating premium coverage of historical product premium duty data of the client according to premium dispersion data, so as to obtain client premium coverage data;

Step S366: and carrying out product coverage evaluation on the historical participating product data of the client according to the client premium coverage data and the client deadline coverage data, thereby obtaining coverage evaluation data.

According to the invention, the characteristics of the insurance product to be evaluated are extracted, so that the time limit and premium distribution condition of the product to be evaluated can be known, and basic data is provided for subsequent analysis. Through the use of customer loyalty data, historical participating products of customers can be screened out, different insurance products purchased by the customers in the past are known, and a data basis is provided for subsequent analysis. The client history product deadline duty ratio data and the deadline coverage data are helpful for evaluating the deadline distribution condition of the insuring products on the client history, determining whether the deadline of the product to be evaluated is matched with the client history policy deadline, and providing preference of the client to products with different deadlines. The customer's historical product premium occupancy and average premium data may be used to learn the premium distribution of the customer's products purchased in the past, as well as the customer's average premium level. This is very helpful in assessing the payment ability and insurance Fei Pianhao of the customer. By analyzing the customer's historical product premium duty and average premium data, the customer's ability to pay and risk tolerance can be assessed. This helps the insurance company to better manage risk, formulate reasonable pricing policies, and ensure sustainability and profitability of the insurance product. By evaluating the historical participating product data and premium coverage data of the customer, the insurance needs and preferences of the customer can be better known, and more personalized and accurate services can be provided for the customer. This helps to establish a long-term stable customer relationship, enhancing customer satisfaction and loyalty. Through evaluating the coverage range of the historical products of the clients, the performance of the products in terms of meeting the demands of the clients can be evaluated, and reference basis is provided for product popularization and marketing. This helps optimize sales strategies and performance assessment systems, improving sales efficiency and success rates for product promotion.

Optionally, step S4 specifically includes:

step S41: extracting payment period characteristics of insurance product data to be evaluated, thereby obtaining product payment period data;

Step S42: carrying out product default interaction frequency calculation according to the product payment period data, thereby obtaining product default interaction frequency data to be detected;

Step S43: carrying out product payment period feature extraction and interaction frequency feature extraction on the insurance product historical data so as to obtain historical product payment period data and historical product actual interaction frequency data;

Step S44: and carrying out customer interaction experience evaluation on the historical product payment period data, the historical product actual interaction frequency data and the default interaction frequency data of the product to be tested, thereby obtaining interaction experience evaluation data.

According to the invention, through extracting the payment period characteristics of the product, the payment period and the payment mode of the product can be known, and further basic data is provided for subsequent interaction frequency calculation and customer interaction experience evaluation. The method is beneficial to evaluating the flexibility and the customizability of the product, and knowing the preference of the customer in payment, thereby providing reference basis for product customization and popularization. The default interaction frequency of the product to be tested is calculated, so that the interaction degree of the customer and the product can be known, and further, a reference is provided for product design and sales strategies. The frequency of use and customer engagement of the product can be evaluated based on the interaction frequency data, providing guidance for product improvement and market location. By extracting the characteristics of the historical product in the payment period and the actual interaction frequency data, the characteristics of the product to be detected can be compared, and the advantages and disadvantages and the market competitiveness of the product can be known. The method is favorable for finding out the performance of the historical product in terms of payment period and interaction frequency, and provides reference comments for product improvement and optimization. Through customer interaction experience evaluation, the attitudes and preferences of customers on the payment period and the interaction frequency of the products can be known, so that guidance is provided for product improvement and marketing. The satisfaction degree and experience feeling of the customers in the actual use process of the product can be evaluated, and improvement suggestions are provided for product design and service promotion.

Optionally, step S44 specifically includes:

step S441: carrying out default interaction frequency calculation according to the historical product payment period data, so as to obtain historical product default interaction frequency data;

step S442: performing difference value finding on the default interaction frequency data of the historical product and the actual interaction frequency data of the historical product, so as to obtain interaction difference value data of the historical product;

step S443: performing difference product data extraction on insurance product historical data according to the historical product interaction difference data, so as to obtain interaction difference product data;

Step S444: performing similarity calculation on the interaction difference product data and the insurance product data to be evaluated so as to obtain product similarity data, and extracting high-similarity product data from the interaction difference product data according to the product similarity data so as to obtain high-similarity product data;

Step S445: carrying out actual interaction frequency prediction on the insurance product data to be evaluated based on the high-similarity product data, thereby obtaining actual interaction frequency prediction data of the product to be evaluated;

step S446: and carrying out default interaction duty ratio calculation according to the default interaction frequency data of the product to be detected and the actual interaction frequency prediction data of the product to be detected, thereby obtaining interaction experience evaluation data.

According to the invention, the default interaction frequency is calculated through the payment period information in the historical data, so that the average customer interaction frequency of the past products is known, and the reference data is provided for comparison and analysis. The method is beneficial to finding out the general trend of the product in the past customer interaction, and provides reference basis for product design and marketing. By calculating the difference between the default interaction frequency and the actual interaction frequency, the change and trend of the historical product in the aspect of customer interaction can be identified, and the mode of actually using the product by the customer can be understood. Interactive difference data for subsequent analysis is provided to help discover the actual needs and desires of the customer for the product. By extracting the interactive difference data of the historical products, a group of product data reflecting the interactive change of the clients can be obtained, and the evolution process of the products and the dynamic property of the clients can be known in depth. Through similarity calculation, the similarity between the historical product and the product to be evaluated can be identified, and references are provided for product positioning, improvement and marketing. The extraction of the high-similarity product data is helpful for finding out historical products with similar customer interaction modes with the products to be evaluated, and provides a basis for subsequent prediction and evaluation. By utilizing the historical interaction data of the high-similarity product, the actual customer interaction frequency of the product to be tested can be predicted, and a basis is provided for expected interaction after the product is pushed out. The method provides possible customer interaction modes of the product to be tested in the future, and is helpful for product design and market strategy formulation. By calculating the duty ratio of the default interaction frequency and the actual interaction frequency, the customer interaction experience of the product can be evaluated to find out whether the customer interacts with the product frequently as expected. The method is beneficial to identifying the advantages and improvement points of the product in the aspect of customer experience, and provides references for product popularization and service promotion.

Optionally, the present specification further provides an evaluation system of an insurance product, for performing an evaluation method of an insurance product as described above, the evaluation system of an insurance product includes:

The risk evaluation model construction module is used for acquiring historical data of the insurance product and carrying out claim settlement risk evaluation on the historical data of the insurance product so as to acquire historical claim settlement risk data; constructing a claim risk evaluation model based on the historical claim risk data;

The claim risk assessment module is used for acquiring the insurance product data to be assessed, and carrying out claim risk assessment on the insurance product data to be assessed through the claim risk assessment model so as to acquire claim risk assessment data;

the coverage evaluation module is used for extracting client features of the historical data of the insurance products so as to obtain historical client data, and classifying the client loyalty of the historical client data so as to obtain the client loyalty data; performing coverage evaluation on insurance product data to be evaluated according to the customer loyalty data, thereby obtaining coverage evaluation data;

The interactive experience evaluation module is used for calculating the default product interaction frequency of the insurance product data to be evaluated so as to obtain default product interaction frequency data to be evaluated, and performing customer interactive experience evaluation on the default product interaction frequency data to be evaluated and the insurance product historical data so as to obtain interactive experience evaluation data;

and the product comprehensive evaluation module is used for comprehensively evaluating the insurance product according to the claim risk evaluation data, the coverage area evaluation data and the interactive experience evaluation data so as to obtain an insurance product evaluation report.

The system for evaluating the insurance product can realize any one of the method for evaluating the insurance product, is used for combining the operation and signal transmission media among all modules to complete the method for evaluating the insurance product, and the internal modules of the system are mutually cooperated, so that the accuracy of the evaluation result of the insurance product is improved.

Drawings

Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of a non-limiting implementation, made with reference to the accompanying drawings in which:

FIG. 1 is a schematic flow chart of the steps of the method for evaluating an insurance product of the present invention;

FIG. 2 is a detailed step flow chart of step S1 of the present invention;

Fig. 3 is a detailed step flow chart of step S3 in the present invention.

The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.

Detailed Description

The following is a clear and complete description of the technical method of the present patent in conjunction with the accompanying drawings, and it is evident that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, are intended to fall within the scope of the present invention.

Furthermore, the drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus a repetitive description thereof will be omitted. Some of the block diagrams shown in the figures are functional entities and do not necessarily correspond to physically or logically separate entities. The functional entities may be implemented in software or in one or more hardware modules or integrated circuits or in different networks and/or processor methods and/or microcontroller methods.

It will be understood that, although the terms "first," "second," etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element, without departing from the scope of example embodiments. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.

In order to achieve the above objective, referring to fig. 1 to 3, the present invention provides a method for evaluating an insurance product, comprising the following steps:

historical data of past insurance products is collected in the embodiment, including applicant information, policy details, claim records and the like. Through careful data analysis, the historical data is subjected to claim settlement risk evaluation by using a statistical method and a machine learning technology, and detailed historical claim settlement risk data is generated. Based on the data, a claim risk evaluation model is established, and the model is established by combining variables such as the age, occupation, health condition and the like of the protected object and using algorithms such as logistic regression or decision trees and the like so as to quantify the claim risk.

In this embodiment, insurance product data to be evaluated including product characteristics, protected object information, and the like is obtained. And evaluating the data through a previously constructed claim risk evaluation model, and calculating the claim risk condition of the product to be tested, namely the claim risk evaluation data. This can help to understand the potential risk level of the product and provide a scientific basis for developing product strategies.

In this embodiment, customer characteristics including purchase frequency, policy renewal, etc. are extracted from insurance product history data. And clustering or classifying the historical customer data to obtain the loyalty data of different customer groups. Based on these data, the potential coverage of the product to be evaluated is evaluated, understanding the market penetration of the product in different customer groups.

In this embodiment, the data of the product to be evaluated is used to calculate the default interaction frequency, that is, the frequency of the interaction between the customer and the product without a specific claim settlement event. And evaluating interactive experience of the product by combining the historical data, wherein the interactive experience comprises indexes such as customer satisfaction, feedback rate and the like, and forming detailed interactive experience evaluation data.

In this embodiment, evaluation data of three aspects of claim risk, coverage and interactive experience are comprehensively considered, and trade-off and comprehensive evaluation are performed. And generating a final insurance product evaluation report by using a proper scoring model such as a weighted total score model, an analytic hierarchy process and the like, providing comprehensive evaluation of products, and providing detailed advice and data support for product improvement, market positioning and popularization strategies.

Optionally, step S1 specifically includes:

step S11: acquiring historical data of insurance products;

in this embodiment, insurance product history data is obtained through an internal database of the insurance company, data provided by a partner, or a public channel. The insurance product history data contains detailed insurance product information including, but not limited to, applicant information, premium, claim records, policy deadlines, and the like.

in this embodiment, a data processing tool is used to extract deadline characteristics from historical data of insurance products. This includes extracting policy terms, payment terms, etc. information from each insurance product. For example, key features such as the insurance period, the start and stop dates of the payment period, and the policy duration of each insurance product may be calculated.

The claims risk assessment is performed in this embodiment using historical data of long-term insurance products. This may include building mathematical models, such as risk models based on historical odds and amounts. Through statistical analysis and machine learning methods, the risk level of claims for long-term products is determined and corresponding evaluation data is generated.

In this embodiment, claims risk assessment is performed on short-term insurance product history data. In a manner similar to long-term products, models can be built to assess the risk of claims to short-term products. This may involve different feature extraction and model parameters to accommodate the specificity of the short-term product.

In the embodiment, the long-term product claim risk data and the short-term product claim risk data are combined, so that the consistency of data formats is ensured. Data cleansing and conversion may be required to enable seamless integration of information of both types of products in subsequent steps.

In this embodiment, a claim risk evaluation model is constructed based on the integrated historical claim risk data. Historical claim risk data may be modeled using methods including statistical modeling, machine learning algorithms, and the like. By training the model, the risk of claims to be resolved for future insurance products can be predicted.

Optionally, step S13 specifically includes:

In this embodiment, by extracting features from the historical long-term insurance product data, key information including claim records and premium payment conditions may be obtained. For the claim data, information such as the number of claims to be paid, the amount of claims to be paid, and the like for each policy may be extracted. And for the premium payment data, information such as the payment frequency, the payment amount and the like of each policy can be extracted. In this way, a historical data set of claims and premium payments for long-term insurance products can be established.

in this embodiment, the historical long-term claim data is used to calculate the probability of claim settlement for the same product. By statistically analyzing these data, the probability distribution of the product claims can be determined and a threshold of claim risk can be set based thereon. This threshold may help to categorize long-term products into high-risk and low-risk categories.

In this embodiment, according to the set risk threshold value of the product claim, the historical long-term claim data can be classified, and the product is classified into two categories, namely, high claim risk and low claim risk. Therefore, the high-claim-risk long-term product data and the low-claim-risk long-term product data can be obtained, and a basis is provided for subsequent risk management and product policy formulation.

in this embodiment, the historical long-term premium payment data is used to count the number of missed payments for each customer. Thus, the conditions of missed high-frequency customer payment and missed low-frequency customer payment can be distinguished. This information can help assess the customer's level of paid credit, thereby better controlling risk.

In the embodiment, the missing payment data of the clients and the historical long-term insurance product data are combined, so that the risks of the clients can be classified. In this way, high-customer-risk long-term product data and low-customer-risk long-term product data can be obtained, and the risk management strategy aiming at different customer groups can be formulated.

In the embodiment, the high risk data of the long-term product for the claim settlement is obtained by performing intersection operation on the long-term product data of the high risk for the claim settlement and the high risk of the customer; likewise, low risk data for long term product claims is obtained by performing an intersection operation on low claim risk and low customer risk long term product data.

In this embodiment, the low risk data of the long-term product claim and the high risk data of the long-term product claim are combined to obtain a complete risk data set of the long-term product claim. This data set may be used in decision making processes for risk management, product design, and pricing.

Optionally, step S14 specifically includes:

In this embodiment, the feature extraction is performed on the historical short-term insurance product data, including information such as customer ID, policy start and stop date, policy, and premium. At the same time, age information of the customer is extracted from these data.

in this embodiment, the historical short term claims data is subjected to claims frequency statistics, which may include counting the number of claims per customer to identify high frequency and low frequency claims clients.

in this embodiment, the customer age data is classified according to the general knowledge of the age of the customer, for example, the customer ages are classified into two groups of the senior age and the senior age.

In this embodiment, the age-limited insurance product is extracted from the historical short-term insurance product data according to the age data of the elderly clients and the age data of the low clients. And then, carrying out intersection operation on the low-frequency short-term claim settlement data and the age-limiting insurance product data, and extracting product data appearing in the two data to obtain low-risk short-term product data, namely short-term product claim settlement low-risk data.

In this embodiment, the client ID in the high-frequency short-term claim data is associated with the client ID in the advanced client age data to acquire first client-associated data. The historical short-term insurance product data is screened for data relating to the customers using the customer IDs in the first customer-associated data. High risk elderly customers, such as customers with ages exceeding a certain threshold and having high frequency claims recorded, are identified according to business rules and risk criteria. Short-term insurance product data, i.e., high risk data for advanced short-term product claims, associated with these customers is extracted.

In this embodiment, the client ID in the high-frequency short-term claim data is associated with the client ID in the low-age client age data to acquire second client-associated data. The data associated with the customers in the historical short-term insurance product data is filtered using the customer IDs in the second customer association data. Such as products that are underwriting and have records of high frequency claims. Relevant short-term insurance product data, namely low-age short-term product claim settlement high risk data, is extracted.

In this embodiment, the low risk data of short-term product claims and the high risk data of advanced short-term product claims are combined. Ensuring that consistent data formats and field identifications are used in merging. The merging operation may be performed using database operations or data processing tools.

Optionally, step S3 specifically includes:

in this embodiment, customer-related information such as customer ID, sex, age, occupation, family condition, etc. is extracted from insurance product history data. Appropriate data processing tools and techniques, such as SQL queries or Python programming, are used to extract the required customer characteristic information from the insurance product history data.

in this embodiment, historical customer data is used to extract product information of customer participation, including product ID, product type, application time, insurance amount, etc. This may be done by querying policy information in the history data or by concatenating the customer data and the policy data.

In this embodiment, statistical analysis is performed on the customer participating product data to distinguish between long-term customers and short-term customers. For each customer, the insurance period of the participating product is calculated, and the insurance period exceeding a certain threshold (such as one year) is defined as a long-term customer, otherwise, the insurance period is defined as a short-term customer. And calculating long-term customer participating product data and short-term customer participating product data, and calculating long-term customer participating product proportion and short-term customer participating product proportion.

In this embodiment, the number of products participating in each customer is counted by counting the data of the products participating in the customer, which may be the total number or the number according to different types of products.

In this embodiment, a customer loyalty division rule is formulated according to the customer participating product proportion data and the number of customer participating products. Clients can be classified into high, medium and low levels of loyalty according to service requirements, and classification rules can be formulated according to analysis of specific data.

In this embodiment, customer loyalty data is used to perform coverage evaluation on insurance product data to be evaluated. The coverage and potential markets of each insurance product in different loyalty customer groups are evaluated based on the characteristics of the customer loyalty in different levels. Statistical analysis or machine learning models may be used for evaluation and prediction.

Optionally, the customer loyalty division in step S35 is specifically:

in this embodiment, the definition of the short-term product is determined according to the guarantee period of the insurance product in the history data (e.g. the guarantee period is less than one year). And analyzing the data of the customer participating products, and calculating the duty ratio of each customer short-term product. Clients with a higher than a certain threshold (e.g., 50%) are identified and classified as short-term product high-duty clients. Similar to short-term products, the definition of long-term products is determined based on the insurance period of the participating products (e.g., the insurance period is greater than one year). And analyzing the data of the customer participating products, and calculating the duty ratio of each customer long-term product. Clients with a higher than a certain threshold are identified and classified as long-term product high-duty clients. Monotonous products refer to products that a customer purchases for only one type (long or short term). The customer participating product data is analyzed, customers participating in only one product are identified, and classified as monotone product customers. And identifying customer groups with the same proportion of the participating products, namely that the proportion of the participating products of a plurality of customers is similar, and no obvious difference exists.

In this embodiment, the number of products per customer is counted. Customers with a number of secured products above a certain threshold (e.g., 3 products) are identified and classified as high secured number customers. The identification of the high-priced participation number customers is similar to the high-priced participation number, but a lower threshold (e.g., 1 or 2 products) is set.

in this embodiment, the intersection operation is performed on the long-term product high-duty customer data and the high-participation-quantity customer data, and the customer data in which both data appear is counted, so as to obtain the first high-loyalty customer data. Customers purchase a relatively large proportion of long-term products, meaning they have confidence in the long-term stable relationship of the insurer, while purchasing multiple products may mean they prefer full coverage guarantee requirements, which are manifestations of loyalty customers, so customers with both high occupancy and high participating amounts are defined as high loyalty customers.

In this embodiment, the short-term product high-duty customer data and the high-participation-quantity customer data are subjected to intersection operation, and customer data in which the two types of data appear equally are counted, so as to obtain second high-loyalty customer data. Customers purchase short-term products more often, meaning their acceptance of corporate products, while purchasing multiple products may mean they prefer full coverage guarantee requirements, which are manifestations of loyalty customers, so customers with both high occupancy and high participation amounts are defined as high loyalty customers.

In this embodiment, intersection operation is performed on the same-duty customer data and the high-participation-number customer data, and customer data in which the two types of data appear equally is counted, so as to obtain third-highest loyalty customer data. The long-term and short-term products purchased by customers are on the same scale, meaning that they are confident in the long-term stable relationship of the insurer and also approve the insurer's products, while purchasing multiple products may mean that they are more prone to full coverage guarantee requirements, which are manifestations of loyalty customers, so customers with the same scale and high number of participants are defined as high loyalty customers.

In this embodiment, intersection operation is performed on monotone product client data and high-amount participating number client data, and client data in which both data appear equally is counted, so that fourth high loyalty client data is obtained, and purchase monotone of long-term products or short-term products represents the trust degree of clients on company products. Long-term product purchasers may be more inclined to establish long-term partnerships, exhibiting higher loyalty. While high-volume participating customers show a comprehensive dependence on the company's products, meaning that they have chosen the company's products in multiple areas of security, which also reflects a degree of loyalty. The intersection operation of the high-participating-amount customer data and the monotone product customer data can screen out customers which not only show loyalty in the amount of products, but also show loyalty in the type of products. Such versatile loyalty manifestations more fully reflect the customer's loyalty level. Considering monotone product customer data or high-priced participation amount customer data alone may have a certain misjudgment, and thus the loyalty of customers cannot be comprehensively and accurately estimated. Through intersection operation, indexes of the two aspects can be mutually supplemented, the risk of misjudgment is reduced, and a high loyalty client is more accurately determined.

In this embodiment, the monotone product customer data and the low-participation-number customer data are subjected to an intersection operation, and the customer data of the same appearance of the two data are counted, so as to obtain low-loyalty customer data, wherein the low-loyalty customer data is compared with those of high-loyalty customers, and the low-loyalty customer data may be characterized by a low proportion of purchased products, a low number of products or lack of overlapping between different loyalty indexes. These customers may not have a high dependency on the company's products or services.

The first high loyalty customer data, the second high loyalty customer data, the third high loyalty customer data, and the fourth high loyalty customer data are combined into a single data set. The low loyalty customer data and the high loyalty customer data are combined into one data set to obtain complete customer loyalty data.

Optionally, step S36 specifically includes:

In this embodiment, feature extraction is performed on insurance product data to be evaluated, and data mining techniques may be used, including, but not limited to, extracting product deadline data and premium data. For example, by analyzing policy information, the term of the insurance product and corresponding premium information are extracted.

in this embodiment, customer history participating product data is extracted from insurance product history data based on customer loyalty data. This may be accomplished by associating the customer ID and screening its historical participation records to obtain a list of insurance products held by the customer.

In this embodiment, product deadline ratio calculation is performed on customer history participating product data. And obtaining the product deadline duty ratio data of the customer history by counting the duty ratio of each product deadline in the customer history. And calculating the deadline coverage by using the deadline data of the product to be evaluated, and reflecting the matching degree of the deadline of the historical participating product of the customer and the product to be evaluated.

In the embodiment, product premium duty ratio calculation is performed on the historical participating product data of the clients, and average premium is calculated. And obtaining the product premium duty ratio data of the customer history by counting the duty ratio of each product premium in the customer history. At the same time, an average premium is calculated as a typical premium level for the customer.

In this embodiment, premium dispersion calculation is performed on premium data of a product to be evaluated and average premium data of customers. This may be a statistical method such as standard deviation, reflecting the degree of premium dispersion. Further, based on the premium dispersion data, a premium coverage of the premium ratio of the customer's historical product is calculated, revealing the comprehensiveness of the premium distribution in the customer's historical product.

In the embodiment, the customer premium coverage data and the deadline coverage data are combined to evaluate the product coverage of the customer historical participating products. The overall performance of the historical participating products of the clients on the matching degree with the products to be evaluated can be evaluated by comprehensively considering the coverage degree of the deadlines and the premium, so that the product coverage evaluation data can be formed.

Optionally, step S4 specifically includes:

In this embodiment, when extracting payment period characteristics of insurance product data to be evaluated, firstly, payment period related data is extracted from policy information. For example, fields such as a payment period, a first payment date, etc. in the policy may be parsed to create product payment period data. By analyzing and sorting the data, the payment period characteristics of the product are formed, including the range of the payment period, the common payment period and the like. For example, the payment period may be monthly, quaternary, annual, etc.

In this embodiment, product default interaction frequency calculation is performed based on product payment period data. And calculating the default interaction frequency of the product by analyzing the specific value of the payment period. For example, if the payment period is monthly, the default interaction frequency may be set to once monthly. Thus, basic data can be provided for subsequent interactive experience evaluation.

In the embodiment, payment period feature extraction and interaction frequency feature extraction are performed on insurance product historical data. And extracting historical payment period data and actual interaction frequency data of the product through policy information in the historical data. This includes information such as the payment period of the history policy, the actual payment date, etc. And forming the payment period characteristics and the actual interaction frequency characteristics of the historical products through the data.

In the embodiment, customer interaction experience evaluation is performed by combining historical product payment period data, historical product actual interaction frequency data and to-be-tested product default interaction frequency data. And comparing the default interaction frequency of the product to be tested with the actual interaction frequency of the historical product to evaluate whether the interaction experience of the product meets the expectations of clients. For example, if the default interaction frequency of the product under test is too high, and the customer is more inclined to the lower actual interaction frequency, this may result in a poor customer experience. The evaluation result can provide guidance for product improvement and form interactive experience evaluation data.

Optionally, step S44 specifically includes:

in this embodiment, the default interaction frequency of the historical product can be calculated through the payment period data of the historical product. For example, if the historical product is paid for once a month, its default interaction frequency is once a month. This can be derived by analyzing the frequency of the payment dates in the historical data, thereby forming default interaction frequency data for the historical product.

In this embodiment, the default interaction frequency data of the historical product is compared with the actual interaction frequency data to obtain a difference value between the default interaction frequency data and the actual interaction frequency data. This can be derived by subtracting the actual interaction frequency from the default interaction frequency of the historical product. Thus, the interaction difference data of the historical product can be obtained to reflect the interaction frequency deviation condition of the historical product.

In this embodiment, the history data of the insurance product is processed by using the interaction difference data of the history product, and the interaction difference product data is extracted. This may be accomplished by applying similar difference calculation methods to apply the interactive differences of the historical products to the historical data of other insurance products to obtain corresponding interactive difference product data.

In this embodiment, product similarity data may be obtained by calculating similarity between the interaction difference product data and the insurance product data to be evaluated. This may employ various similarity calculation methods, such as cosine similarity or euclidean distance, etc. And then selecting a product with high similarity as a reference according to the similarity data to obtain high-similarity product data.

In this embodiment, based on the high-similarity product data, actual interaction frequency prediction is performed on the insurance product data to be evaluated. The actual interaction frequency of the high-similarity product can be applied to the product to be evaluated, so that the actual interaction frequency prediction data of the product to be evaluated can be obtained.

In this embodiment, a default interaction duty ratio is calculated according to default interaction frequency data and actual interaction frequency prediction data of a product to be tested. The interactive experience evaluation data can be obtained by comparing the difference between the actual interaction frequency predicted value of the product to be tested and the default interaction frequency value and then comparing the difference with the default interaction frequency value.

Optionally, the present specification further provides an evaluation system of an insurance product for performing the evaluation method of an insurance product as described above, the evaluation system of an insurance product comprising:

The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

The foregoing is only a specific embodiment of the invention to enable those skilled in the art to understand or practice the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method of evaluating an insurance product, comprising the steps of:

2. The method for evaluating an insurance product according to claim 1, wherein step S1 is specifically:

step S11: acquiring historical data of insurance products;

3. The method for evaluating an insurance product according to claim 2, wherein step S13 is specifically:

4. A method of evaluating an insurance product according to claim 3, wherein step S14 is specifically:

5. The method for evaluating an insurance product according to claim 1, wherein step S3 is specifically:

6. The method for evaluating an insurance product according to claim 5, wherein the customer loyalty program in step S35 is specifically:

7. The method for evaluating an insurance product according to claim 5, wherein step S36 is specifically:

8. The method for evaluating an insurance product according to claim 1, wherein step S4 is specifically:

9. The method for evaluating an insurance product according to claim 8, wherein step S44 is specifically:

10. An insurance product evaluation system for performing an insurance product evaluation method according to claim 1, the insurance product evaluation system comprising: