Customer Defection Prediction in Online Bookstores

Customer classification based on the probability of customer churn was an important content of customer relationship management. The process of customer churn was analyzed by the logistic regression model. And this paper built a customer churn prediction model based on the analysis of customer ’s transaction data and the behavior pattern. Taking a domestic branch of China Construction Bank as an example, the paper forecast the defection of this commercial bank’s VIP customers. And we can find that the method has a high accuracy rate and covering rate, and provide an effective measurement for customer classification.