Semantic Search in Portals using Ontologies

Portals have become very popular on the internet. They allow users to access many online transactions and customized information. More recently, semantic portals emerged as a natural extension of traditional portals and they have been attracting researchers and enterprise attention as a new way to provide information. Ontologies are the basis of semantic portals, and one of their major functions is to categorize information inside portals. However, there are few practical applications showing effectively how and where ontologies can be used in this environment. This paper covers these issues and shows how ontologies can improve results relevance in one of the most important component of a portal: the search mechanism. It describes a search tool developed in the context of a semantic portal named PASS (Portal with Access to Semantic Search). This tool uses components of domain ontologies stored in the portal to expand terms during the search process. Its evaluation has been done using traditional unities of measure such as precision and recall, and average precision. Their corresponding evaluation graphics showed that searches performed according to our strategy returned more significant results than those generated by a traditional Web search mechanism. 1