Rating the Dominance of Concepts in Semantic Taxonomies

The descriptive concepts of “semantic” taxonomies are assigned to content items the publishing domain for supporting a plethora operations, mostly regarding organization and discoverability content, as well recommendation tasks. However, either not all publishers rely on such structures, or in many cases employ their own proprietary taxonomies, thus is difficult be retrieved by end users stored publisher-specific fragmented “data-silos”, respectively. To address these issues, modular scalable “Dominance Metric” methodology proposed rating dominance importance semantic taxonomies. Our metric applied both vast multidisciplinary Microsoft Academic Graph Fields Study taxonomy MeSH controlled vocabulary order enhanced refined versions produced. Moreover, we describe cleansing process resulting from Microsoft’s structure deduplicating refining hierarchical relations towards increase its representation quality. evaluation procedure provided valuable insights showcasing that high volume, namely number publications concept to, does necessarily imply influence, but latter also affected structural topological properties individual entities.