Collaborative filtering recommendation algorithm based on user interest evolution

Abstract

Personalized recommendation systems provide personalized item recommendations during a live user interaction, and they have achieved widespread success in electronic commerce nowadays. In many personalized recommender systems, collaborative filtering algorithm is the most famous technique and especially in collaborative filtering methods, neighborhood formation is an essential algorithm component. In order to make a recommendation in collaborative filtering algorithm, it is required to form a set of users sharing similar interests to the target user. But traditional collaborative filtering recommendation algorithm does not consider the evolution of user interest when finding the nearest neighbors in different time periods. And the recommendation results can not reflect the user's true interests. For this reason, a personalized collaborative filtering recommendation algorithm based on user interest evolution is given. This recommendation approach takes into account the important factor that user interests changes over time. © 2011 Springer-Verlag Berlin Heidelberg.