Detection Of Substructures In Telecom Social Network Community For Churn Prediction

Abstract

Conceptual In the telecom segment, a tremendous volume of information is being produced every day because of a huge customer base. The telecom companies consider that accomplishing new customer is costlier than holding the current ones. This paper proposes a social network analysis based model used to identify the positive key customers. The proposed approach builds a telecom social network; it explores basic network features, and identifies the community structures present in the network. The most important part of substructure mining is clique structures which share social structure, usage structures and common interests. The clique is a group of customers in the telecom social network, they share similar properties. The paper identifies positive key customers, its clique subgroup and that substructure is used to identify the churn in telecom customer’s network.