Exploiting relation extraction for ontology alignment

摘要: When multiple ontologies are used within one application system, aligning the is a prerequisite for interoperability and unhampered semantic navigation search. Various methods have been proposed to compute mappings between elements from different ontologies, majority of which being based on various kinds similarity measures. As major shortcoming these it difficult decode semantics results achieved. In addition, in many cases they miss important due poorly developed ontology structures or dissimilar designs. I propose complementary approach making massive use relation extraction techniques applied broad-coverage text corpora. This able detect types relations, dependent used. Furthermore, exploiting external background knowledge, can relations even without clear evidence input themselves.