Metadata and semantics: A case study on semantic searching in web system

Metadata is used in Web information systems to improve search results. Searching for resources based on metadata usually relies on keyword matching, i.e. user's query terms must match with metadata terms of the relevant resources. The Semantic Web framework incorporates ontology with metadata to retrieve resources. This paper describes a case study of semantically enhanced searches in a Web information system. The study combines ontologies with metadata to augment searches in a Web resource collection. Ontologies representing semantic associations between the US presidents were created and allow some semantic queries for the domain. For example, a semantic query for resources about "the US president who is the successor of Bill Clinton" can be formed instead of the query term "George W. Bush". The semantic queries were formed using description logic expressions. Evaluation of the results demonstrated good precision with moderate recall. Factors contributing to and detracting from retrieval performance were identified and addressed. The results highlight both the potential and issues in combining ontologies with metadata for enhancing search in next-generation Web information systems.