INTEGRATING WEB PORTALS INTO A CONCEPT-BASED SEARCH ENGINE USING ONTOLOGIES
Xueqiang Wang · 2004
Concept-based searching provides a promising approach for exploiting in-formation from the Internet. It is especially suitable for the next generation Web, the semantic Web. It can also be used to retrieve information from the Deep Webs, where the data is stored in searchable database hidden behind portals. One challenge here is to integrate these portals and provide a uni-fied query interface for all data sources of certain domain. In this thesis the author presented an approach to integrate web portals into a concept-based search engine. The main idea is to match user query dynamically with query interfaces of search engines by consulting the knowledge in a domain ontol-ogy. Both user queries and query interfaces are represented based on the universal schema in the ontology. Then they can be easily matched to gener-ate queries. From users ’ perspective, there is a unified query interface. Be-sides, the algorithm COLD is introduced and implemented in this thesis. It is an attempt to categorize search engines by analyzing the closeness be-tween the attributes in each search engine. Finally, the author presented how