Analysis on Forecasting the Consumption of Telecom Customers -Study Based on the Kunming Data
In this paper we extract 18, 950 cost standard customers from the original data of 29, 999 telecom customers from Kunming city and various counties, and make the time series model and logistic regression model, hoping to judge the customers online situation and test the accuracy of the models respectively through the amount of consumption of the customers from the previous months. The purpose of this paper is to provide early warnings for telecom companies against freak customers whose charges and quantity of phone calls drop dramatically, so as to take relevant measures to let them stay on the line effectively.
telecom customer drain time series model logistic regression model
ZHAO Jinjin FEI Yu
School of Statistics and Mathematics, Yunnan University of Finance and Economics, P.R.China, 650221
国际会议
威海
英文
448-452
2010-07-24（万方平台首次上网日期，不代表论文的发表时间）