Exploiting rich side information sources of user and items for cross domain collaborative recommendation

Cross-domain collaborative filtering (CDCF) aims to alleviate the sparsity problem in individual CF domains by transferring knowledge among related domains. The core concept of CDCF is to exploit information from multiple User-Item (U-I) matrices (i.e. domains) in order to allow the recommendation performance of one domain to benefit from the information from one or more other domains. In other words, we can regard CDCF as Collaborative Filtering on one U-I matrix/domain that takes other U-I matrices as additional information sources. In this paper, we will give a brief survey of the pilot studies in this research line in two dimensions: Rich information sources for collaborative filtering and Knowledge Transfer Styles.