TWM563034U - Insurance underwriting system - Google Patents

Abstract

An insurance underwriting system comprising an underwritting risk module is disclosed. The underwritting risk module, constructed with a plurality of descriptive variables classified into several categories, is configured to perform at least one statistic algorithm with the descriptive variables to generate one single risk variable for risk evaluation on underwriting an insurance application. When an insurance application is submitted to the insurance underwriting system, the insurance underwriting system accesses information of the corresponding information to the descriptive variables, puts the information into the statistic algorithm, and generate a value of the risk variable as a score of underwriting risk of the insurance application.

Description

Insurance underwriting system

This creation is related to an insurance underwriting system, and especially to an insurance underwriting system applied to an insurance industry.

In the past, the operational procedures of underwriting insurance relied on manual operations, which comprehensively considered various conditions and information of the insured and evaluated their risks to decide whether to approve the insurance. Therefore, the underwriting operation time usually takes several days, and the manpower is limited to increase efficiency, and may be missed due to the negligence of manual operations. The above-mentioned problems have caused various costs increase for the insurance industry over the years, and it is urgent to find a solution or improvement.

One of the purposes of this creation is to provide an insurance underwriting system by setting up an underwriting risk model, interpreting variables in it, and executing at least one statistical analysis algorithm to generate a single risk variable for underwriting risk assessment of an insurance case.

One of the purposes of this creation is to provide an insurance underwriting system that uses the underwriting risk model to bring insurer data corresponding to the insured case into the statistical analysis algorithm to generate a single risk variable value as one of the insured cases. Underwriting risk score.

According to this creation, an insurance underwriting system is provided, including an underwriting risk model. The underwriting risk model includes a plurality of explanatory variables classified into several variable categories, and is set to generate a single risk variable by interpreting the variables and executing at least one statistical analysis algorithm to perform an underwriting risk assessment on an insurance case. When the insurance case is put into the underwriting risk model for underwriting risk assessment, the underwriting risk model obtains the insured data corresponding to the insured case through interpretation variables, and brings the corresponding insured data into the statistical analysis algorithm to generate a single risk variable The value is used as an underwriting risk score for one of the insurance cases.

From the above, it can be known that the insurance underwriting system of this creation automatically generates underwriting risk scores for insurance cases through underwriting risk models, which can reduce manual work and improve the efficiency and correctness of underwriting.

In order to further explain the embodiments, the present invention provides drawings. These drawings are part of the content of this creative disclosure. They are mainly used to explain the embodiment, and can be used to explain the operation principle of the embodiment in conjunction with the related description in the description. With reference to these contents, those skilled in the art should be able to understand other possible implementations and the advantages of this creation. Elements in the figures are not drawn to scale, and similar element symbols are often used to indicate similar elements.

In order to develop the technology of automatic underwriting, the creator uses appropriate algorithms to gradually build corresponding risk prediction models according to risk categories (such as various types of insurance contracts such as hospitalization, surgery, total disability, major illness, etc.) as the The basis of intelligent underwriting.

According to an embodiment, an underwriting system is provided, including an underwriting risk model. The underwriting risk model includes a plurality of explanatory variables classified into several variable categories, and is set to generate a single risk variable by interpreting the variables and executing at least one statistical analysis algorithm to perform an underwriting risk assessment on an insurance case. When the insurance case is put into the underwriting risk model for underwriting risk assessment, the underwriting risk model obtains the insured data corresponding to the insured case through interpretation variables, and brings the corresponding insured data into the statistical analysis algorithm to generate a single risk variable The value is used as an underwriting risk score for one of the insurance cases.

According to another embodiment, an underwriting method is provided, which includes the following steps: establishing an underwriting risk model, which includes a plurality of explanatory variables classified into several variable categories, and the underwriting risk model is set through the explanatory variables and Executing at least one statistical analysis algorithm to generate a single risk variable for underwriting risk assessment of an insured case; and when an insured case is put into an underwriting risk model for underwriting risk assessment, the underwriting risk model is used to obtain insurance through interpretation variables The data of the insured person corresponding to the case, and the corresponding insured person information is brought into the statistical analysis algorithm to generate a single risk variable value, which is used as an underwriting risk score for one of the insurance cases.

Please refer to FIG. 1 and FIG. 2 together. FIG. 1 shows a system architecture diagram of an underwriting system 1 according to an embodiment of the present invention, and FIG. 2 shows a process of an underwriting method 100 according to an embodiment of the present invention. Illustration. Please note that for the sake of simplicity, the underwriting system 1 and the underwriting method 100 are described together. However, in other embodiments, the underwriting method need not be limited to the underwriting system 1 illustrated in FIG. 1. The underwriting system 1 includes an input module 11, an underwriting risk model 12, and an evaluation module 13. The input module 11 can receive data related to an insurance case invested by the front-end system (not shown), and then transmit the data to the underwriting risk model 12 and the evaluation module 13. The evaluation module 13 controls the underwriting risk model 12 to carry out the insurance case. Underwriting risk assessment . The underwriting risk model 12 includes a number of explanatory variables that are classified into several variable categories. The establishment of the underwriting risk model 12 (step 110) can be divided into two stages. The first is the stage of the field construction, combining the experience of the original human reviewer, confirming the target variables, and developing the corresponding possible risk factors, and considering The availability of data, aggregated multiple explanatory variables for subsequent model building. The second stage is the model construction stage. In order to optimize the benefits of the underwriting risk model 12 at this time, the machine learning (Maching Learning) method can be selectively applied, and the risk factors hidden in the data (Risk Factors) are used as explanatory variables. .

For example, in this embodiment, there are more than 160 explanatory variables in eight types of variable categories, but the types of variables and the number of explanatory variables are not limited to this. The eight variable categories include, but are not limited to: personal data categories, insured data categories, self-generated schedule categories, historical insurance record categories, policy loan categories, handler categories, claims information categories, and family information categories. Examples of explanatory variables included in the personal data category include, but are not limited to, gender, age, height, weight, BMI value, occupation, history of hospitalization, history of hospitalization, past illness, physical disability and so on. Examples of explanatory variables included in the data category of the book to be secured include, but are not limited to, payment classification, relationship with the insured, whether it is a collection, county, city, and so on. Examples of explanatory variables included in the spontaneous adjustment category include, but are not limited to: marital status, smoking, drinking, betel nut, source of contract, purpose of insurance, main financial source, work vehicle, education of insured person, insured People's education, annual income to be insured, source of income, etc. Examples of explanatory variables included in the historical insurance record categories include, but are not limited to: the total number of main contracted policies insured in the past five years, the total number of contracted insurance policies in the past five years, the total number of valid main contract policies, the total number of valid contracted insurance policies, and whether they are new and many more. Examples of explanatory variables included in the policy loan category include, but are not limited to: the current policy loan amount, total policy loan amount, and so on. Examples of explanatory variables included in the category of handlers include, but are not limited to, the relationship between the handler and the insured, the type of reconciliation, the number of contracts cancelled within two years, and so on. Examples of explanatory variables included in the claim data category include: whether claims have been claimed in the past five years, whether family members have claimed anti-cancer insurance, whether family members have claimed major illness insurance, and so on. Examples of explanatory variables included in the family data category include, but are not limited to, the number of family members, the number of family members insured against cancer, the number of family members covered by major illness insurance, and so on.

Secondly, the underwriting risk model 12 is set to generate a single risk variable by interpreting variables and executing at least one statistical analysis algorithm to perform an underwriting risk assessment on an insurance case. Here, Principal Components Analysis is used as an example of a statistical analysis algorithm. However, in other embodiments, other types of statistical analysis algorithms, such as a regression analysis algorithm, may be applied, and there is no limitation to this. Principal component analysis can reduce the number of explanatory variables (X) to the coordinates of the target variable (Y) and reduce the number of variables, but at the same time keep the uniqueness of each variable. The underwriting risk score calculated by the statistical analysis algorithm can represent the value of the risk probability of the insured case incurring a claim within a predetermined time. Therefore, after the introduction of underwriting risk model 12, new insurance cases can automatically calculate risk scores to assist in early warning of those with higher risks in the underwriting of insurance cases, and automatic underwriting of those with lower risks, which is expected to increase. Underwriting efficiency, reducing short-term risks and other benefits.

Secondly, in step 120, when the underwriting risk model 12 receives an insurance case via the input module 11, the insurance account information corresponding to the insurance case is obtained through the interpretation variable, and various insurance account information is brought into the corresponding interpretation variables. The front-end system is exemplified but not limited to a database, an insurance contract case system, or other electronic devices or systems that can obtain information about an insurance case. The underwriting risk model 12 brings the above-mentioned corresponding insured data into a statistical analysis algorithm to generate a single risk variable value as an underwriting risk score for one of the insurance cases.

For example, when the insured case is related to a medical insurance, the underwriting risk model 12 can use statistical analysis algorithms to convert the explanatory variables into a relatively small number of variables and regression to generate a single risk variable. In this example, the five variables may include the number of operations, the number of operations, the number of days in the hospital, the number of hospitalizations, the number of visits, and so on.

In another example, when the insured case is related to a medical insurance, the underwriting risk model 12 can use statistical analysis algorithms to convert explanatory variables into a single risk variable, such as any of the death variable and the full residual variable. Or a combination.

Thereafter, optionally, at step 130, an automatic underwriting procedure for a predetermined insurance category is performed, including: when the underwriting risk score is higher than a risk threshold, submitting an insurance case for manual review, and underwriting When the risk score is lower than the risk threshold, the case is automatically underwritten.

Therefore, from the above, it can be known that through the underwriting system and its method to automatically generate the underwriting risk score of the insured case through the underwriting risk model, it can reduce manual work and improve the efficiency and correctness of underwriting.

The above description creates a number of different embodiments according to the present invention, wherein each feature can be implemented in a single or different combination. Therefore, the disclosure of the implementation mode of this creation is to clarify the specific embodiment of the principle of the creation, and the invention should not be limited to the disclosed embodiment. Furthermore, the previous description and its drawings are only for the purpose of this creative demonstration and are not limited by it. Changes or combinations of other elements are possible without departing from the spirit and scope of this creation.

1‧‧‧ Underwriting System
11‧‧‧input module
12‧‧‧Underwriting Risk Model
13‧‧‧ Evaluation Module
100‧‧‧ Underwriting method
110, 120, 130‧‧‧ steps

The drawings of the present invention are described as follows: FIG. 1 shows a system architecture diagram of an underwriting system according to an embodiment of the present invention; and FIG. 2 shows a flowchart of an underwriting method according to an embodiment of the present invention.

Claims (9)

An insurance underwriting system includes: an underwriting risk model including a plurality of explanatory variables classified into several variable categories, and the underwriting risk model is set to be generated through the explanatory variables and executing at least one statistical analysis algorithm A single risk variable for underwriting risk assessment of an insured case; where, when the insured case is put into the underwriting risk model for underwriting risk assessment, the underwriting risk model: obtains the insured case through the explanatory variables Corresponding policyholder data, and bringing these corresponding policyholder data into the statistical analysis algorithm, generate a single risk variable value as an underwriting risk score for one of the insurance cases. The insurance underwriting system described in item 1 of the scope of patent application, wherein when the insured case is related to a medical insurance, the underwriting risk model converts these explanatory variables into five variables by a principal component analysis method. A single risk variable is generated. The five variables include the number of operations, the number of operations, the number of days in the hospital, the number of hospitalizations, and the number of visits. The insurance underwriting system described in item 1 of the scope of patent application, wherein when the insured case is related to a medical insurance, the underwriting risk model converts these explanatory variables into a single risk variable through the statistical analysis algorithm, This risk variable includes the death variable and the total residual variable. The insurance underwriting system described in item 1 of the scope of patent application, wherein the underwriting risk model is further set to perform an automatic underwriting procedure for a predetermined insurance category, including: where the underwriting risk score is above a risk threshold At that time, the insurance case is submitted for manual review, and when the underwriting risk score is lower than the risk threshold, the insurance case is automatically underwritten. The insurance underwriting system according to item 1 of the scope of patent application, wherein the statistical analysis algorithm includes a regression analysis algorithm, and the underwriting risk score represents a risk probability value of the insurance case for claim settlement within a predetermined time. The insurance underwriting system according to item 4 of the scope of patent application, wherein the predetermined time is the next four years. The insurance underwriting system as described in item 1 of the scope of patent application, wherein the variable categories include personal data categories, types of insured materials, self-generated adjustment categories, historical insurance record categories, policy loan categories, handler categories, Claim information category and family information category. The insurance underwriting system described in item 1 of the scope of patent application, wherein the explanatory variables include gender, age, height, weight, BMI value, occupation type, history of illness, history of hospitalization, past illness, physical disability, payment, To be insured relationship, whether it is a collection, county, city, marital status, smoking, drinking, betel nut, source of contract, purpose of insurance, main financial source, work vehicle, education of the insured, insured Education, annual income of the insured, source of income, total number of main contracted insurance policies in the past five years, total number of insured contracted insurance policies in the past five years, total number of valid main contracted insurance policies, total number of valid contracted insurance policies, whether new insurance holders, current Amount of policy loan, total amount of policy loan, relationship between handler and insured, type of adjustment, number of contracts cancelled within two years, whether claims have been claimed in the past five years, whether family members have paid for cancer insurance, or whether family members have paid Major illness insurance, number of families, number of family anti-cancer insurance applicants, number of family critical illness insurance applicants. According to the insurance underwriting system described in item 1 of the scope of patent application, it further includes an input module to receive data related to an insurance case, and an evaluation module to control the underwriting risk model to evaluate the underwriting risk of the insurance case.