Research On Auto Insurance Renewal Based On Machine Learning Method

Since its emergence,the insurance industry has played an important role in the national economic development system and has irreplaceable important value in the social construction system and the government management system.The insurance industry benefit from the process of the market economy gradually in-depth has become one of the fastest-growing industries in the national economy system.With the large increase in disposable income,car prices fell sharply and the gradual improvement of residents purchase environment,China’s auto market expanding,auto demand and has the high speed growth trends.The first big property insurance coverage is the car insurance.In 2019,the percent of all property insurance premium income Sixty-one is the income of auto insurance premium.Like other types of insurance,only when insurance companies attract new customers and pay attention to the renewal status of old customers can auto insurance occupy a larger market share.Therefore,it is very necessary for insurance companies to study and predict the renewal status of auto insurance.In this paper,we use Logistic regression model,the decision tree model,the random forest model of car insurance renewal case study,the three models are all of the classic model of machine learning to deal with the problem of classification and prediction because of simple operation,good effect of fitting,high prediction accuracy and is widely used.This paper on the three classical machine learning model are introduced,including the definition,distribution,parameter estimation model and building steps,the algorithm of the model,etc.Then,the use of original data and the data processing method for this article,including the data contained in the information,the characteristics of the variables,etc.,as well as the data missing value Principle,data grouping,integration methods.At the end of the paper has established the Logistic regression forecasting model,the decision tree model,random forest prediction model,using respectively the three kinds of prediction model for a certain insurance company car insurance renewal situation forecast analysis.Comparing the results of these three models,the prediction accuracy of three models of Logistic model respectively seventy-one point seven percent eighty point four percent seventy-five point four percent,the decision tree model,the random forest model,which can predict results of the effect of random forest model the best conclusion.Car insurance renewal model research has important significance and widely There is room for further study and improvement of the three models used in this paper.