Research On Auto Insurance Renewal Rate Based On Kernel Logistic Regression

Recently,the international insurance industry has developed steadily,and auto insurance accounts for the majority of property and casualty insurance expenses.And with the introduction of commercial fee reforms related to auto insurance,auto insurance consumers have been given more choices.Therefore,in order to win the market and increase the insurance coverage rate,various property companies often take preferential measures to compete for customers,which at the same time causes a decline in profit margins.Therefore,in order to improve the profit margin of property insurance companies is very important.Because new customers need higher costs,it is necessary to pay attention to the issue of renewal of old customers.Auto insurance renewal means the maintenance of old customers,while auto insurance non-renewal business is an expansion of the business.In contrast,getting new customers requires more investment for distance.Therefore,it is very necessary to study the issue of auto insurance renewal.The above is the background of this article,using the customer renewal data of an insurance company as the research object,based on the kernel logistic regression model from the perspective of practical application to establish the auto insurance renewal model,and research and analyze the factors that affect the auto insurance renewal rate.First of all,to understand the business knowledge of auto insurance renewal,this article analyzes and understands the situation of auto insurance renewal at home and abroad,including the research on auto insurance abroad and the explanation of the influencing factors and research methods of auto insurance renewal in China.Through a certain understanding,I know the importance of establishing an effective,reasonable and efficient auto insurance renewal model.Secondly,this article describes the knowledge of the renewal business,determines the purpose of the business,defines auto insurance renewal customers and auto insurance non-renewable customers,introduces the methods of data collection and cleaning,variable screening,and model construction and test.Then,based on the auto insurance renewal model established by the kernel logistic regression model,the auto insurance renewal data in this paper is used to conduct empirical research on the auto insurance renewal data.Based on the business knowledge,the data is descriptive statistical analysis,and the missing values and abnormalities in the data are analyzed.The value is processed.When selecting the variables into the model,we analyze them according to the correlation and feature importance of the variables,and filter the variables in and out of the model.Finally,the sorted data is divided into a training data set and a verification data set,and the data set is trained using the method of control variables to select the optimal solution of the relevant parameters of the model.The training results show that the established auto insurance renewal model obtains higher AUC and accuracy evaluation values,which verifies the effectiveness of the model.Aiming at the results of the experiment,this article also analyzes the factors that affect auto insurance renewal to provide help for practical applications.