Web Search Disambiguation by Collaborative Tagging

Download

	PDF (Paper) Download (233Kb)
	PDF (Presentation File) Download (1108Kb)

Description/Abstract

Existing Web search engines such as Google mostly adopt a keyword-based approach, which matches the keywords in a query submitted by a user with the keywords characterising the indexed Web documents, and is quite successful in general in helping users locate useful documents. However, when the keyword submitted by the user is ambiguous, the search result usually consists of documents related to various meanings of the keyword, in which probably only one of them is interesting to the user. In this paper we attempt to provide a solution to this problem by using the semantics extracted from collaborative tagging in the social bookmarking site del.icio.us. For an ambiguous word, we extract sets of tags which are related to it in different contexts by performing a community-discovery algorithm on folksonomy networks. The sets of tags are then used to disambiguate search results returned by del.icio.us and Google. Experimental results show that our method is able to disambiguate the documents returned by the two systems with high precision.