Enhancing a Biological Concept Ontology to Fuzzy Relational Ontology with Relations Mined from Text

In this paper we investigate the problem of enriching an existing biological concept ontology into a fuzzy relational ontology structure using generic biological relations and their strengths mined from tagged biological text documents. Though biological relations in a text are defined between a pair of entities, the entities are usually tagged by their concept names in a tagged corpus. Since the tags themselves are related taxonomically, as given in the ontology, the mined relations have to be properly characterized before entering them into the ontology. We have proposed a mechanism to generalize each relation to be defined at the most appropriate level of specificity, before it can be added to the ontology. Since the mined relations have varying degrees of associations with various biological concepts, an appropriate fuzzy membership generation mechanism is proposed to fuzzify the strengths of the relations. Extensive experimentation has been conducted over the entire GENIA corpus and the results of enhancing the GENIA ontology are presented in the paper.