The Application Of Data Mining In Telecom Customers Churn

With the development of data mining technology, people become more and more interested in its application in various fields. Data mining is to find out the hidden business rules through the establishment of the model by using historical data. The telecom industry is a typical data-intensive industry, and currently telecom operators are suffering different degrees of customers churn. Therefore, it is particularly important to use data mining to identify the characteristics of the churn customer.The purpose of this paper is to detect the characteristics of the lost customer, establish churn prediction model, establish the lost customer list by using the model, and makes an effective detainment strategy against the churn customers.The main contents of this paper include the following aspects: Firstly, from the business view, this paper analyses the business process of telecom industry and the reason of customer churn. Then this paper combines the reason of customer churn and the telecom business system, achieving the conversion from commerce view to data view. Secondly, from the technical view, this paper summarizes the theory and method of data mining and describes the algorithm of neural network and decision tree in detail. Then, this paper describes the principles of establishing the model of the telecom industry customer churn, and discusses the standard data mining process. This paper describes the six stages of CRISP-DM standard data mining process, and discusses the business understanding, data understanding, data preparation, model building, model evaluation, and model implement. In the end, this paper discusses the value of retaining customers and the characters of all stages of the customer life cycle. Then this paper analyzes the reason of customer churn and makes an effective detainment strategy against the churn customers.