Building Data Mining Applications For CRM

This article aims to improve network education services and analyzes in particular the characteristics of such services. Based on comprehensive usage of customer lifecycle ideas in customer relationship management, targeted subjects are thereby divided into two groups: potential learners and present learners. For potential learners, to better understand their needs on the program, difficulties in click-stream date handling and means of attribute-oriented induction are described heavily in this article, which is to help predict the learners'program interests. For present learners, however, weaknesses of network education supporting system are to be demonstrated and outlined in this article. At the same time, an S-tutor supporting system model based on the web is established, putting forward a new model of learning sources and that of users, coming up with an orientation for dynamic studying. Through which clustering analysis of data mining is focused on and a cooperative studying system verified. Ultimately, a solution of collecting click-stream data on user's level is actualized with specific examples, and the potential learners'program interests are analyzed. Meanwhile, a contrast K-means and hierarchical clustering computation is made. This research topic has obtained warm support from "MIS of Xidian Univeristy Network Education School".This article represents a summary and improvement of such relative theories.