The Application Of Data Mining In Telecom Customer Loss Predictions

Over the past ten years,the international and domestic communications market has undergone great changes,and the popularization of 4G network has sped up the market reshuffle.The current domestic communication market is saturated.For any operator who wants to increase new customers,it becomes increasingly more difficult,and the cost becomes increasingly higher.How to keep existing customers becomes more and more important.As the market becomes more open and people's requirement on number portability becomes stronger,operators are attaching more and more attentions to the problem of user loss nowadays.Even if user loss is 2% per month,the company will suffer huge loss of profit.Customer loss is a very big problem faced by operators.It is necessary to implement an effective customer retention plan.In this study,a model was created to detect user loss,and determine the most important factor which causes customer loss.The purpose of this study is to develop a system model to predict customer loss.It would be the first step in the company's implementation of retention plan to predict which customers will be lost.We can use the system to generate a customer list,so that we could implement personalized retention plan accordingly.In this article,predecessors' r esearch were summarized,actual data was taken as the basis and data mining technologies such as decision trees,neural networks,etc.were adopted.Thus,a customer loss prediction model was established so as to effectively explore lost customers,and ret ain them.Data mining contains a large number of technologies,and it is applied in the analyzing the possibility of customer loss.By changing classifiers used in modeling and conducting a large number of simulations,the best solution can be found among them.When conducting data mining,the standard data mining process CRISP-DM was adopted in this study.