Extending an on-line information site with accurate domain-dependent extracts from the World Wide Web

This paper describes a new procedure that has been developed for extending an existing on-line information system about The Voyages of the Beagle with information collected automatically from Internet. A Term Identification procedure finds relevant terms in the document; and the algorithm uses conventional search engines (such as Google) to look for pages about those terms. Next, a sequence of filters rule out all the information considered irrelevant, and the remaining data is put together in “summary pages” available to the students. Our experiments so far have attained very good results, and in a form that was sent to several users of the on-line site they all showed much excitement about the tool.

• Type: Conference
• Year: 2003
• Where: CAISE
• Authors: Enrique Alfonseca, Pilar Rodríguez
• Keywords: CAISE 2003, Conventional Search Engines, Relevant Terms, Term Identification Procedure