Enabling semantic similarity estimation across multiple ontologies: An evaluation in the biomedical domain

► Methods to enable the estimation of semantic similarity between different ontologies are presented. ► They rely on the assessment of the semantic overlapping and structural similarity between ontological concepts. ► Evaluated with several edge-counting similarity measures, two biomedical benchmarks and MeSH and WordNet as ontologies. ► Results show a noticeable improvement in similarity accuracy compared to approaches based on terminological matchings.