Collaborative Recommendation Algorithm Based on Semi-Supervised Learning
Siqi Jiang, Yufeng Liu, Yuxin Zhou, Huanhuan Zhi · 2017
Data sparseness is one of the key issues existed in the collaborative filtering recommendation system. In this paper, we propose a novel algorithm named Collaborative Recommendation algorithm Based on Semi-Supervised Learning (SSLCF). First, build heterogeneous information networks through combine multi-dimensional information such as users, items, labels. Second, we can use similarity (level of interest) as weight between isomorphism (heterogeneous) nodes. Second, we use regularization framework algorithm to discriminate label information for unlabeled users and items, we predict rating and generate recommendation results according to the preferences category of the target users. Experimental results show that SSLCF significantly outperforms the state-of-theart methods. The results shows the proposed model can solve the few label data issue and helps to improve the quality of recommendation.