Research on E-Commerce Recommendation Service Using Collaborative Filtering

Currently collaborative filtering is the most successful and widely used recommendation technology in e-commerce recommender systems. The idea behind this technology is that it may be of benefit to user´ s search for products by examining the behavior of other users who share the same or similar interests with her/him. In this paper, an e-commerce recommender system using collaborative filtering, called ECRec, is proposed. ECRec is designed and realized on the client/server architecture, including four function modules (RecDB, RecEngine, configuration console and Monitor Agnet). Moreover, ECRec employs two basic CF algorithms and four improved CF algorithms for different situations. Due to its independence to business system of e-commerce Web site, ECRec has better portability, maintainability, and the characteristics of open architecture.