Discovering the Context of WWW Pages to Improve the Effectiveness of Local Search Engines

摘要

This work proposes a method of searching for information in hypertext systems representing WWW sites. The method is based on the creation of a 2-level index. The first level of the index is related to information located only inside the nodes. The second level of the index relates to information which is not restricted to one node but encompasses a set of related nodes. The second level is based on the context hierarchy which is a hierarchical organization of the main themes dealt with by the information contained in the site and gives a notion of context to the pages. This notion permits a new operator named context: to be added to the query language allowing the user to better express his information need.