AN IMPROVED COLLABORATIVE FILTERING METHOD FOR PRODUCT RECOMMENDATION BASED ON CUSTOMER TRANSACTION DATA

摘要

Recommender systems are widely used by E-commerce sites to suggest products to theircustomers. In this paper, an improved collaborative filtering methodbased on customer transaction datais proposed to improve the quality of recommendations. Firstly, integrated RFM (IRFM) indicators are computed to evaluate each customer's preferences based on his purchase history, then an improved collaborative filtering method is used to provide personalized recommendation to target customer according to the user-item IRFM matrix. This recommendation method overcomes the limitations of traditional collaborative filtering methods and is able to provide high quality recommendation to active customers.