An Efficient Method of Building the Telecom Social Network for Churn Prediction

As the deregulation and great advance of new technologies, the competition in wireless telecommunication industry is getting severe. Hence, churn prediction and management have become of great concern to themobile operators. Therefore they wish to retain their subscribers and satisfy their needs. Previous methods address the homogeneous social network analysis for churn prediction by considering the single relation. From the point of view of data mining, a social network is a dynamic, heterogeneous and multirelational in nature. Typical work on social network analysis includes the construction of multi-relational telecommunication social network and discovery of group of customers who share similar properties and classify the customers as churners and non-churners. In this paper we explore the various methods of representing the social networks. Considering the multi-relational data while constructing the telecomsocial network will increase the efficiency in prediction of customer churning.