Prediction Of Customer Churn In Insurance Industry Based On Data Mining

The basic idea of data mining is to use known data to establish mathematical models to find hidden laws.In the insurance industry,the cost of acquiring new customers is much higher than that of retaining existing customers.From the perspective of the future,good data mining applications will be of great help to the development of business operations and the discovery of potential customer groups.Therefore,the integration of data mining technology and commercial enterprises is also an inevitable process.Due to the continuous improvement of office automation level,the informatization of insurance companies,and the continuous renewal of the concept of corporate governance,the traditional and high-quality insurance relationship management methods with customers no longer meet the basic needs of modern insurance companies.The main reasons for the loss of insurance enterprises' customers are: their own problems,price problems and quality problems.However,we should not blindly reduce the price.For example,for price sensitive customers,reducing the price will keep the price insensitive customers,and adjusting the price can not retain them,but also reduce the profit margin.Therefore,we should understand the trigger events of customer churn.Due to the increasingly mature data extraction technology,such as classification method and other data extraction methods are also increasingly mature,the application of such technology to the analysis of insured customer information,customer information into a number of useful information,which ultimately helps enterprises to make business decisions,while reducing the loss rate of customers,so as to obtain business advantages.In particular,customer churn analysis is an important part of insurance company analysis.This paper analyzes the historical data of customer churn.At present,we are studying the characteristics of customer delivery and expect to take appropriate measures to reduce customer churn.This is very important to reduce the transaction cost of insurance companies and improve their business efficiency.This study first explains the meaning,function and process of data mining,analyzes the meaning and system framework of customer relationship management and its application in insurance industry,and analyzes the main reasons of customer churn of insurance companies.Then the advantages and disadvantages of decision tree algorithm are analyzed.In this paper,we improve the algorithm and propose a decision tree classification mining algorithm with weight attribute and pre cleaning strategy.The improved algorithm is more efficient and can deal with a large number of data.Establish and analyze the algorithm of the new model,collect commercial insurance companies,apply the customer churn prediction model established in this paper,clean up the data contained in the application process of the model,convert the data,and establish the customer churn prediction model through decision tree C5.0 and C &RT,and evaluate and analyze the results of the prediction model.