Recommending entities with similar types is an important part of entity recommendation, particularly for multi-type entities. So there is a necessity to measure similarity between multi-type entities. However, most existing similarity measures are simply based on either type collection intersection or type vector similarity, and pay little attention to the weighting of types. In this paper, we propose an EMD-based similarity measure for multi-type entities, which not only takes into account pairwise type similarity, but also the weighting of types. We also present a novel PageRank-based weighting scheme by using type hierarchy. The experimental results show that our weighting scheme outperforms base-line weighting schemes and that our EMD-based similarity measure outperforms traditional similarity measures. © 2014 Springer International Publishing Switzerland.
CITATION STYLE
Zheng, L., & Qu, Y. (2014). An EMD-based similarity measure for multi-type entities using type hierarchy. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8709 LNCS, pp. 282–293). Springer Verlag. https://doi.org/10.1007/978-3-319-11116-2_25
Mendeley helps you to discover research relevant for your work.