Combining Linked Open Data Similarity and Relatedness for Cross OSN Recommendation

Abstract

The emergence of online social networks (OSNs) and linked open data (LOD) bring up opportunities to experiment on a new generation of cross-domain recommender systems in which the true benefit of LOD can be exploited, particularly to address the new user problems. In this article, the authors explore the feasibility of combining the two axes of comparison, similarity and relatedness, in LOD space, and introduce a new LOD-based similarity measure. The reason is to take benefit more from LOD to compare general resources, which can be useful in the context of cross-OSN recommendation. Experimental evaluation demonstrates the effectiveness of the proposed approach.