A New Suffix Tree Similarity Measure and Labeling for Web Search Results Clustering

Due to the enormous size of the web and low precision of user queries, finding the right information from the web can be difficult if not impossible. One approach that tries to solve this problem is using clustering techniques for grouping similar documents together in order to facilitate presentation of results in more compact form and enable thematic browsing of the results set. Web search results clustering is an attempt to apply the idea of clustering to document references (snippets) returned by a search engine in response to a query. Thus, it can be perceived as a way of organizing the snippets into set of meaningful thematic groups. This paper introduces a new similarity criterion for merging which is evaluated for search results returned from actual web search engines.