Research On The Telecom Customer Churn Prediction System Based On Data Mining

Under the condition of market economy, customers are the source of competitive advantages. Customer-centric concept is not only a slogan, but also a strategic choice of the enterprise. Maintaining old customers and developing new customers are often used in customer management. How to maintain old customers and prevent customer churn are the concern problem for many industries, especially in the highly competitive telecom market. In order to improve the competitiveness of enterprise, firstly predicting effectively the situation of potential loss of customers could be solved, and appropriate measures are taken to reduce losses, and achieve maximum benefits. With the development of data mining technology, the research based on data mining prediction becomes more widely. Data mining techniques has the processing capability of mass data and sophisticated data mining algorithms that make it play a huge role in the telecom customer churn prediction.This thesis introduced research background and significance of data mining technology used in telecom customer churn prediction, and reviewed the modeling methods of customer churn prediction. The customer business data in a telecom company were used, the CRISP-DM methodology became the modeling process framework, SPSS Clementine as data mining tools was adopted, to build the index system of customer churn prediction based on the integrated algorithms of decision trees and neural network, and established the predicting model of customer churn. Finally, the system of predicting customer churn was built based on B/S structure, using the C#programming language, SQL Server2008database and ASP.NET technology.The system provides a reference for telecom marketers to early identify losing customers, make a viable customer retention programs, reduce the customer churn, and strengthen telecom customer churn management. Meanwhile, it has a wide range of industrial applications prospects and practical value for the relevant enterprises to reduce operating costs and improve operational performance.