Use of Logistics Regression Model in Credit Evaluation for Mobile Subscribers

摘要

As mobile communication market appears to be saturated, the customer relationship management has become a key element for revenue stabilization for mobile communication network operators. In order to reduce bad debt, operators often set credit limits for subscribers, but for those customers who are not defaulting intentionally, the incorrect credit limit will result to decreasing customer satisfaction and losing revenue directly. In this paper, logistics regression model is applied to resolve the problem. At first we select effective possible explanatory variables from those possible variables based on the information about customer behavior through chisquare test and contingency test, and then establish the credit evaluation model for mobile subscribers include 6 variables, at last we verify the efficiency and stability of the model by comparison test. With the help of this model deployed in IT system, operators can not only assign credit ratings and set credit limits for subscribers automatically, but also enhance customer satisfaction and improve profit margins.