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

China's telecommunications industry has experienced an unprecedented rapid development,the market share tends to saturation,the operator's profit is gradually declining.In the face of this trend,the importance of customer retention is particularly prominent.Telecom operators from their own point of view,customers of telecom industry operators,if can use the scientific method to discover from the network tendency of users,and find out the reasons for the loss of their customers,according to the actual situation and take measures to develop specific implementation,will more effectively retain customers in order to effectively control the loss of customers.With the development of science and technology,the data is increasing,the performance of the computer has been improved,the data mining technology has been widely used.Data mining can be used to clean up the data,synthesis,processing,and then modeling and analysis,not only can transfer information between the data,but also to use historical data for future classification and prediction.The data is used in customer data of telecom enterprises,the use of data mining techniques,the use of data mining commercial software SPSS Modeler,through the analysis of the prediction of useful information in data mining for the loss of customers,to find out the factors causing the loss of customers,comprehensive use of neural network and decision tree algorithm,a loss forecasting model of customer,and in order to construct the system of customer loss index weight,and the use of Jave language on the basis of the above,JSP technology and SQL server database,finally developed based on B/S structure prediction system.More intuitive showing the probability of loss of customers and scores.Telecom operators through the loss of customer early warning system to use,although in the preliminary forecast work has played a certain effect,to a large extent to bring considerable benefits to the enterprise.But due to time constraints and limitations,and the limitations of the author level,there is further room for improvement and perfection,the subject finally build prediction model for the management of customer churn follow-up in telecommunication enterprises,provide a valuable reference.