The web search engine works on the World Wide Web upon human command. In addition to advances in the young science of information retrieval, development of the search engine is confined by the evolution and understanding of both of the Web and the human mind, despite the satisfying performance achieved so far, e.g. by Google. As a self-contained thesis, it aims to extract concepts from the web pages returned from the search engine and organise the concepts in a conceptual graph. Each of the retrieved concepts is treated as a query and input to the search engine and more concepts will be extracted from the result pages. Thus, a graph connecting concepts, called a Concept Relation Network (CRN), is obtained. This thesis studies several ways to weight the connections between concepts in CR...[ Read more ]

The web search engine works on the World Wide Web upon human command. In addition to advances in the young science of information retrieval, development of the search engine is confined by the evolution and understanding of both of the Web and the human mind, despite the satisfying performance achieved so far, e.g. by Google. As a self-contained thesis, it aims to extract concepts from the web pages returned from the search engine and organise the concepts in a conceptual graph. Each of the retrieved concepts is treated as a query and input to the search engine and more concepts will be extracted from the result pages. Thus, a graph connecting concepts, called a Concept Relation Network (CRN), is obtained. This thesis studies several ways to weight the connections between concepts in CRN. Analysis has been performed to identify whether a concept concerns a physical location or not and to concept clusters bearing certain graph topology. Finally, the application of CRN to search engine personalisation is studied.