An ontology mapping framework
- Pathum Thani, Thailand : Asian Institute of Technology, 2003
- 57 p.
- Thesis ; no. CS-03-22 .
- Asian Institute of Technology. Thesis ; no. CS-03-22 .
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering, School of Advanced Technologies
Thesis (M.Eng.) - Asian Institute of Technology, 2003
The Semantic Web becomes ever more important issue for the Web today. It relies strongly on formal ontologies to structure data for comprehensive and exchangeable machine understanding. Furthermore, information on the Web is naturally distributed. Each information source uses its own ontology to represent its data. An interoperability problem among different information sources is therefore a challenge issue on the Semantic Web. One way to overcome that problem is to provide a systematic way for mapping and merging ontologies. However, there are some limitations for recent related research. Most approaches are usually based on heuristic or syntactic clues to determine equivalence between ontologies. Thus, a new ontology mapping framework is proposed by applying a quantitative technique called semantic similarity measure to compute a similarity value between two concepts. The semantic similarity measure can be retrieved from properly combining three semantic factors consisting of the shortest path length between two concepts, the depth and the information content of a subsumer by using WordNet, an online lexical dictionary, as a semantic reference source. Besides, a new concept is also introduced, called the weight-based approach, to support strong semantic relation word pairs by adding weight into semantic similarity measures. Consequently, this technique displays a dramatic improvement over the previous research. Based on the proposed approach, the automatic financial report mapping system is developed as a demonstration system of its practical usage in a real-word domain. iii