UKSearch - Web Search with Knowledge-Rich Indices
Udo Kruschwitz · 2000
Standard search engines prove to be very useful if they have access to huge amounts of data. However, a common problem is search over a restricted domain. This paper addresses this problem by indexing the source data in a more elaborate way than in standard search engine technology. This allows us to extract concepts that are used to create a structure for the documents that is similar to that found in classified directories. Overview The vast amount of data that can be found on the Web helps to find information on specific topics very quickly. However, a common problem is to find the appropriate documents in well defined subdomains of the Web. For example the search for lecturers in AI on the Essex university Web pages is not successful even though the information is there, but not explicitly on a single page. A related problem is the retrieval of too many documents: searching for lecturers gives us too many matches in an apparently random order. Our approach aims at creati...