A Novel K-medoids clustering recommendation algorithm based on probability distribution for collaborative filtering

作者：

Highlights：

• An improved Kullback–Leibler (KL) divergence is introduced to calculate item similarity.

• The optimum center of our algorithm is obtained by maximizing the contribution sum of distance.

• An asymmetric mode is used to emphasize the asymmetric relationship between items.

• Results show that our scheme improves the effectiveness of recommendation systems.

摘要

•An improved Kullback–Leibler (KL) divergence is introduced to calculate item similarity.•The optimum center of our algorithm is obtained by maximizing the contribution sum of distance.•An asymmetric mode is used to emphasize the asymmetric relationship between items.•Results show that our scheme improves the effectiveness of recommendation systems.

论文关键词：Collaborative filtering,Clustering,K-medoids algorithm,Kullback–Leibler (KL) divergence,Probability distribution

论文评审过程：Received 28 September 2018, Revised 6 March 2019, Accepted 15 March 2019, Available online 21 March 2019, Version of Record 26 April 2019.

论文官网地址：https://doi.org/10.1016/j.knosys.2019.03.009