There are two significant impedances to the realization of the potential of ontologies. First, many ontology designers lack the necessary background in formal logics to express their intuitions clearly and precisely, resulting in the proliferation of ontologies with low expressivity. Concurrently, developing semantic mappings between existing ontologies is difficult, because much of the semantics is external to the representation. This thesis uses the idea of metaphor to develop architectures for ontology repositories to serve as bottom-up reusable resources. Moreover, an ontology design algorithm has been developed that allows designers to communicate their ideas at the semantic level, simply by generating and vetting models. Finally, a semantic mapping algorithm has been developed that uses an ontology repository to determine the similarities and differences between any number of target ontologies. An ontology for partial orders has been elaborated to demonstrate the proof of concept and populate the first iteration of the repository.
Identifer | oai:union.ndltd.org:TORONTO/oai:tspace.library.utoronto.ca:1807/17512 |
Date | 10 August 2009 |
Creators | Hashemi, Ali |
Contributors | Gruninger, Michael John |
Source Sets | University of Toronto |
Language | en_ca |
Detected Language | English |
Type | Thesis |
Page generated in 0.1522 seconds