A novel recommendation method based on social network using matrix factorization technique

作者：

Highlights：

• We propose a hybrid clustering algorithm which composes of KHM and PSO.

• Our hybrid clustering algorithm overcomes the sensitivity of initial conditions.

• We consider many factors in the process of similarity computation.

• We use matrix factorization technique to alleviate the data sparsity and cold-start problems.

摘要

•We propose a hybrid clustering algorithm which composes of KHM and PSO.•Our hybrid clustering algorithm overcomes the sensitivity of initial conditions.•We consider many factors in the process of similarity computation.•We use matrix factorization technique to alleviate the data sparsity and cold-start problems.

论文关键词：Recommendation method,Social network,K-harmonic means,Particle swarm optimization,Matrix factorization

论文评审过程：Received 15 April 2017, Revised 26 December 2017, Accepted 21 February 2018, Available online 27 February 2018, Version of Record 27 February 2018.

论文官网地址：https://doi.org/10.1016/j.ipm.2018.02.005