An efficient collaborative recommendation algorithm based on item clustering

Abstract

To help people to find useful information efficiently and effectively, information filtering technique emerges as the times require. Collaborative recommendation is becoming a popular one, but traditional collaborative recommendation algorithm has the problem of sparsity, which will influence the efficiency of prediction. Unfortunately, with the tremendous growth in the amount of items and users, the lack of original rating poses some key challenges for recommendation quality. Aiming at the problem of data sparsity for recommender systems, an efficient collaborative recommendation algorithm based on item clustering is presented. This method uses the item clustering technology to fill the vacant ratings where necessary at first, then uses collaborative recommendation to form nearest neighborhood, and lastly generates recommendations. The collaborative recommendation based on item clustering smoothing can alleviate the sparsity issue in collaborative recommendation algorithms. © 2010 Springer-Verlag Berlin Heidelberg.