The Prediction Model Study Of 3G Customer Churn Based On Data Mining Of China Unicom

The 4th-generation(4G)mobile license of China were issued in December4,2013,it means that the telecom operation market will face to a new pattern of full competition among three operators(China Unicom,China Mobile & China Telecom).Under the circumstances of the low penetration of 4G and the near saturation of telecom market,3G customer is particularly important.Spurred by the increasing competition,the rising risk of 3G customer loss prompt operators ponder over how to complete the program of 3G customer churn prediction and customer sustaining and retaining.Combined with the 3G customer churn prediction program of a certain branch of China Unicom,the project stated for the theory of Data Mining,utilized diversified algorithms to set up the model of customer churn prediction,and select the C5.0 as the main algorithm by comparing the classification precision and recognition efficiency.From the overall situation,this essay firstly according to the definitely state for the and for its customer loss,and carry on the analysis to the erosion status of the branch from the 3G customer and customer quality loss and the loss of contracted users;And then based on the CRISP-DM(gross-industry process for data mining method of data mining theory,according to the steps of understanding of business,data explanation,data preparation,modeling,set up the churn prediction model of 3G customers in October,and provides a method for discovering the loss-customers and providing the matching dimension in measures and the process of marketing and maintaining.The essay also predicts the 3G customer churn in December by calculate the stability.The ending of the essay,its afford the instructions of the model and do a prospect of the future work.