A Fast Interactive Item-Based Collaborative Filtering Algorithm

Abstract

A recommender system becomes more and more popular in e-commerce. Usually prediction results cannot satisfy users’ requirements fully, and sometimes it even contains totally irrelevant items. To reflect users’ newest preference and increase the quality of recommendation, a fast interactive item-based collaborative filtering algorithm is proposed. Firstly, we propose an item-based collaborative filtering algorithm with less time and space complexity. Then we introduce interactive iterations to reflect users’ up-to-date preference and increase users’ satisfaction. The experiments show that our fast interactive item-based CF algorithm has better recall and precision than traditional item-based CF algorithm.