This dissertation studied various means of calculating similarity in the annotations of web pages compared to the similarity of the document text. A software tool, named Semantic Web Analysis of Similarity (SW AS), was developed and utilized to perform the analysis of similarity in annotated documents published from the first three years of the International Semantic Web Conference. Rules concerning the ontological concepts of the documents were specified and these rules as well as other parameters were varied to determine the effect on overall similarity measures. Traditional measures of similarity as well as enhanced measures of similarity proposed for this study were evaluated. A proposal was made concerning use of similarity measures to evaluate the consistency of semantic annotation for documents published through a Semantic Web portal.
| Date of Award | Jan 1 2008 |
|---|
| Original language | English |
|---|
| Supervisor | Junping Sun (Supervisor), James D. Canandy (Advisor) & Michael J Laszlo (Advisor) |
|---|