Semantic Relatedness In RDF Knowledge Graphs

Published: 14 September 2021| Version 2 | DOI: 10.17632/xws5zf5y9j.2
Contributor:
Francesco Taglino

Description

These data refer to the benchmarking of 10 methods for computing semantic similarity applied to 13 golden datasets. The 10 addressed methods are: - Wikipedia Link-based Measure (WLM ) [Witten and Milne, 2008]; - Linked Open Data Description Overlap-based approach (LODDO) [Zhou et al. 2012] - Linked Data Semantic Distance (LDSD) [Passant, 2010] - Linked Data Semantic Distance with Global Normalization (PLDSDGN) [Piao and Breslin, 2016] - Propagated Linked Data Semantic Distance (PLDSD) [Alfarhood et al. 2017] - Information Content-based approach [Schuhmacher and Ponzetto, 2014] - REWOrD [Pirrò, 2012] - Exclusivity-based [Hulpuş et al 2015] - ASRMP [El vaigh et al. 2020] - Proximity-based [Leal, 2013] The 13 golden datasets are: Atlasify240, B0, B1, GM30, MTurk287, R122, RG65, MC30, KORE-IT, KORE-HW, KORE-VG, KORE-TV, KORE-CN In the experimentation, two DBpedia knowledge grapsh have been considered, i.e., with and without the dbo:wikiPageWikiLink links.

Files