Abstract

Ontologies are going to play a significant role in the future semantic web. They help organize the information about domains to facilitate not only to retrieve information about the domain but also reasoning and derivation of new information which can be used by search engines and software agents. This major report discusses the resources offered by ontologies and explores the way to use them in web document detection, annotation and search. Since an ontology captures all the significant terms and their hierarchical relationships for a specific domain, it motivates us to use this information to detect web documents which fall under that domain. The content of the web documents can be analyzed to find out the number of ontology terms encountered in those documents and accordingly assign them rank indicating how close they are to the corresponding domain. Once we know the encountered ontology terms in the web document, we can annotate the terms with appropriate web links where more information about the terms can be found. The domain detection approach can also be very useful in searching web documents. If we know the search domain in addition to search terms, the search can be more effective and accurate. While searching the web, we usually get search results containing documents from different domains whereas we are interested in the documents of a particular domain. Domain detection technique using ontologies might solve this problem. In this report, the algorithm of domain detection, annotation and search for web documents exploiting the resources offered by ontologies has been explained and implemented as a Java application. The report also discusses the design of a test ontology that has been used to test the implemented system.