meX-Search is a meta search engine that automatically categorizes search results into thematic groups and displays them by intuitive and interactive maps.

meX-search is an experimental, non commercial meta search engine built up from april to july 2004 by Karsten Knorr during his diploma thesis in computer and media science [University of Applied Science Berlin]. The main idea of the thesis was the implementation of an intuitive and simple user interface for web clustering search engines.

Users of conventional Web search engines are often forced to sift through a long list of off-topic documents to find relevant results… Especially when the search query is general, it is often hard to find relevant resources among thousands of irrelevant ones. Search result clustering is a approach to handle such problems by grouping similar documents among search results into thematic groups.

meX is a meta search engine. Currently meX is getting the search results completely from the Yahoo-API.

The clustering of the result-snippets from Yahoo is based on Carrot2, an open source java framework for clustering textual data. Within the Carrot2 framework meX uses the Lingo Algorithm. The Authors of the Carrot2 framework and components: Dawid Weiss, Jerzy Stefanowski, Stanislaw Osinski.

Tags: clustering, google, graphical user interface, information retrieval, information visualization, maps, search engine