Novel clustering algorithm for moderating the risk of customer churn

Abstract As market competition surging everyday in the telecom sector, customer churn management has become an imperative for the telecom organizations to enhance their profit levels and provide better services. The conventional churn prediction models in the telecom sector don’t work well while dealing with the big data. Decision makers are consistently confronted with inaccurate operations management. As a solution to these adversities, several clustering methods are proposed such as Semantic Driven Subtractive Clustering Method (SDSCM) with K-means and K-median algorithms which are failed to address processing of numerous amount of data. The proposed system is a method of implementing Semantic Driven Subtractive Clustering Method (SDSCM) with K-medoid algorithm, which is capable of processing gigantic data sets and provides 3-D Trajectories that help in efficient decision making by simulating marketing strategies to ensure profit maximization. Keywords: Customer churn, Clustering, K-medoids, SDSCM, Map Reduce.