Data sparsity is a common issue in recommendation systems, particularly collaborative filtering. In real recommendation scenarios, user preferences are often quantitatively sparse because of the application nature. To address the issue, we proposed a knowledge graph-based semantic information enhancement mechanism to enrich the user preferences. Specifically, the proposed Hierarchical Collaborative Embedding (HCE) model leverages both network structure and text info embedded in knowledge bases to supplement traditional collaborative filtering. The HCE model jointly learns the latent representations from user preferences, linkages between items and knowledge base, as well as the semantic representations from knowledge base. Experiment results on GitHub dataset demonstrated that semantic information from knowledge base has been properly captured, resulting improved recommendation performance. Copyright © 2018 Springer International Publishing AG, part of Springer Nature.
|Title of host publication
|Advances in knowledge discovery and data mining: 22nd Pacific-Asia Conference, PAKDD 2018, Melbourne, VIC, Australia, June 3-6, 2018, proceedings, part II
|Dinh PHUNG, Vincent S. TSENG, Geoffrey I. WEBB, Bao HO, Mohadeseh GANJI, Lida RASHIDI
|Place of Publication
|Published - 2018