Today's web search engines need to combine search results from many independent sources, such as results from a Web crawl, from Advertisements, from Video's, Books, News, Shopping, etc. Many advanced features of such search engines might be seen as "sources" too, such as query autocompletion, related queries, spelling correction ("Did you mean:"), and instant answers: These features receive a query and return some data (a completed query, a related query, a corrected query, etc.) Searsia provides an approach to search that:

  • Manages and shares large collections of independent sources.
  • Selects for each query the most relevant sources.
  • Combines sources in an aggregated search interface.
  • Learns over time what kind of information each source provides.

Searsia is demonstrated by UTwente Search, an enterprise search engine for the University of Twente that combines the results of about 35 sources, including results from Google's web crawl, from Courses, from News, the Telephone directory, the Timetables, as well as results from social media, such as Facebook, Twitter and Flickr.

Search by Searsia

University of Twente