Telecommunications user behaviors analysis based on fuzzy C-means clustering

Abstract

As the number of telecommunication user close to saturation, operators turn their focus from how to increase the subscriber to how to maintain the existing ones, which need more in-depth analysis of user character. In this paper we modeling user communication behavior based on the incoming/outgoing call holding time and then use fuzz c-means clustering algorithm to classify every level in user pyramidal model. For each level we get 3 classifications. We analyze the proportion and communication trend of each classification to help operators know their subscribers better. The method and conclusion of this paper can be used as the base of precision marketing for telecommunications industry. © 2010 Springer-Verlag Berlin Heidelberg.