Benchmarking Ontologies: Bigger or Better?

摘要: A scientific ontology is a formal representation of knowledge within domain, typically including central concepts, their properties, and relations. With the rise computers high-throughput data collection, ontologies have become essential to mining sharing across communities in biomedical sciences. Powerful approaches exist for testing internal consistency an ontology, but not assessing fidelity its domain representation. We introduce family metrics that describe breadth depth with which represents domain. then test these using (1) four most common medical respect corpus documents (2) seven popular English thesauri three corpora sample language from medicine, news, novels. Here we show our approach captures quality ontological guides efforts narrow breach between collective discourse Our results also demonstrate key features ontologies, thesauri, different domains. Medical small intersection, as do thesauri. Moreover, dialects characteristic distinct domains vary strikingly many same words are used quite differently As intended mirror state knowledge, methods tighten fit will increase relevance new areas science improve accuracy power inferences computed them.