The Research And Application Of Data Mining On Telecommunication CRM System

As China's three major operators of various services rapidly improved,but also brought fierce competition,and customer satisfaction,loyalty,awareness and other issues of increasing concern.How to improve customer value and retain attract customers,how to make the operation of enterprises will be the real goal of enterprise transformation to the customer as the center of the business model,is the basic motive for many enterprises to pursue greater commercial interests,become the core problem concerned with enterprise decision makersAiming at the application of telecom CRM system,this paper makes a deep research on customer churn prediction,customer satisfaction research and parallelization algorithm research,and the main work is as follows:(1)based on decision tree and GSO algorithm,the churn prediction model is improved to deal with large numbers of non lost instances and unbalanced classes of telecom data sets,resulting in imbalanced training sets of clusters.Experimental results show that the algorithm can effectively predict customer churn;(2)based on the model of EFQM and ECSI,an improved model for analyzing customer satisfaction is put forward.The experimental results of the improved model show that the algorithm is more efficient in improving customer satisfaction index;(3)use the MapReduce model of Hadoop platform to implement parallel calculation of the improved prediction method,and realize the parallel calculation of the customer satisfaction model analysis method,so as to improve the operation efficiency.