Clustering web search results using Suffix Tree Clustering.
Search engines like Google display web-search results in a linear order. Normally, a query word like "jaguar" would result in pages that contain information on cars and cats and any other "jaguar"s out there. Searching would be a much better experience for the user if the results were clustered under these topics (which are automatically extracted from the results page). A search engine cannot expect a user to know before hand that there are many categories of Jaguars and she has to refine her search. But if the results page suggests that there are many clusters and actually catergorizes the results into these clusters, the user can search better.
Required Reading:
Oren Zamir and Oren Etzioni, Web Document Clustering: A Feasibility Demonstration, SIGIR 98, Melbourne, Australia.