ICBCF: One item-classification-based collaborative filtering algorithm
Zilei Sun, Nianlong Luo, Wei Hua Kuang · 2011
With the development of personalized recommendation, recommendation algorithms usually need to consider the specific feature of the system so as to obtain more information and get a better result. To improve the regular collaborative filtering algorithms, which is inefficiency and less concerned about item classification, this paper proposes a new item-classification-based algorithm. It proposes the concept of “User Interest Vector”, in order to present users interests and rating tendency better, and then correct the classification information of all the items. We believe this algorithm, which has a better accuracy and lower computation complexity in experiments, is worth popularization and becoming a new research direction of collaborative filtering algorithm.