Research On Telecom Customer Churn Prediction Based On Data Mining

Classification is a very popular topic in data mining research. Given a set of data instances, each belonging to one of a number of predefined classes, the goal is to discover rules that can allow records with unknown class membership to be correctly classified. In this regard, numerous pattern discovery algorithms based on machine learning techniques have been developed leading to effective and efficient classification. Data mining techniques target various applications such as churn analysis, cross-selling, fraud detection, risk management, customer segmentation, sales forecast, etc.Nowadays, churn analysis has become critical for companies involved in the fast changing and fiercely competitive telecommunication market. Application of data mining technologies to customer churn prediction applications enables effective prediction of the likelihood of customers turning to another service provider. This helps service providers improve customer relationship management which in long run benefits the enterprise.In relation to the telecom’s customer churn dataset characteristics, this paper put forward a decision tree method of data mining technology to pinpoint customer churn characteristics and some of the rules relative to customer churn. Via this, the carrier can choose to provide a special personalized offer and services to those subscribers having higher probability of leaving the network in near future. The experiment of customer churn analysis on decision tree is illustrated deliberately. It provides a new research idea and analysis methodology with early warning characteristics for preventing customers from turning over and giving the telecom operator or service provider an opportunity to further enhance its competitive edge. This paper also focuses on efficiency and accuracy of decision tree method to develop a model that successfully handles such classification scenarios.