earticle

영어

Considering the probability matrix decomposition model, direct trust and indirect trust, trust propagation, as well as similarity between users and other factors, this paper proposes a new recommendation algorithm which fully taps the trust mechanism, and takes into account the similarity between users. Firstly, a new trust matrix is obtained based on the trust value of social network. Then all the trust relationships are integrated into the recommendation algorithm through the RSTE model. Finally, the prediction score is produced and the recommended list is given. Experimental results on public data sets show that this recommendation algorithm has obvious feasibility and superiority.