Email Users Churn Analysis Based on PMCLP and Decision Tree

The Web sites often attract people to be their customers, and email is one important tool to deal with that. Thus, the numbers of email users for one website is often an index to show the attraction of people to one website. In this issue, the churn analysis is essential for the website to predict and describe who will leave or give up the use of the email. In this paper, Email users churn analysis is predicted and described by PMCLP (penalized multi-criteria linear programming) and decision tree method. In data mining and machine learning, PMCLP is a classification method with penalized multi-criteria linear programming and Decision tree is often used to deal with classification problem with a descriptive outcome. The first method shows its performance in the prediction result and the second method shows its performance in the predict result description.