Is Ontology Alignment like Analogy? -- Knowledge Integration with LISA

Elie Raad (1) , Joerg Evermann (1)

1 Faculty Of Business Administration

Ontologies are formal descriptions of a domain. With the growth of the semantic web, an increasing number of related ontologies with overlapping domain coverage are available. Their integration requires ontology alignment, a determination of which concepts in a source ontology are like concepts in a target ontology. This paper presents a novel approach to this problem by applying analogical reasoning, an area of cognitive science that has seen much recent work, to the ontology alignment problem. We investigate the performance of the LISA cognitive analogy algorithm and present results that show its performance relative to other algorithms.

Mots clés: Ontology; Alignment; Analogy; LISA; Cognition

Domaines: Recherche d'information [cs.IR]; Informatique

Elie Raad, Joerg Evermann. Is Ontology Alignment like Analogy? -- Knowledge Integration with LISA. Symposium On Applied Computing (SAC), Mar 2014, South Korea. pp.294-301. ⟨hal-00967738⟩