|
With the further reform of telecommunications , communications, market competition has been very intense. Customers are our most important asset is the focus of market competition , customer retention bottom line of the company's profits have amazing effects , far more than company size , market share, unit costs , and many other competitive advantages are usually considered relevant factors . Especially for Telecom, a global telecommunications company almost every being or will be established customer churn model , there is no loss or ready to build the model company is uncompetitive . Data mining is the information industry in recent years, one of the hotspots , data mining historical data from a large number of them in advance to find unknown implicit knowledge , this paper studies the loss of customers , the use of a decision tree BP neural network and other data mining techniques, by analyzing the factors affecting customer churn and the relationship between these factors and found that knowledge , and thus the loss of customers to predict the future . This paper starts from the second chapter , introduced the current data mining technology development status and level , including the definition of data mining , the main technical , architecture, methodology and some specific algorithm , the main aim of a comprehensive analysis of the general sense from the current mainstream data mining techniques . Chapter III combines the latest churn management theory, from the perspective of software engineering, requirements analysis , the overall system architecture design ideas, module structure. The fourth chapter focuses on the entire process of data mining , specifically the establishment of a large customer churn models , and through the results of the implementation of the model for tracking and analysis. Finally , the author summarizes the work done and outlook , noting that during the study facing major problems and difficulties, and further analytical work made ??a number of recommendations.
|