On Customer Churn and Early Warning Model of Telecom Broadband

Abstract

Telecom broadband is a main channel supporting internet surfing in China. With the market competition development, customer churn management has become a kernel task of marketing for telecommunication operators. The traditional market research methods are difficult to support the challenge of churn. Data mining techniques are applied to the customer churn management, to establish an early-warning model for this non-steady-state customer system. The data mining process makes use of C5.0, Logistics regression, and neural network algorithm to train telecom broadband customer dataset in the Pearl River Delta, involving mainly customer demographic data, non-satisfaction complaint, and transition cost. The customer character and assessing model are also discussed.