September 10, 1999
Lehrstuhl für Wirtschaftsinformatik, Aachen University of Technology, GERMANY
Every time a user issues a query he tells us something about his actual information needs. Probably he wants to reuse the queries he executed some time ago maybe with a light topic shift. The query he had issued once probably would be a good starting point for him. Concerning a sequence of queries we will have sets of queries in this sequence that belong together because they are dealing with the same topic. Usually, the queries which are topically related will not be issued in one contiguous perpetual stream. If the user wants to reuse a query that is part of that stream he has to recover it. For the recovery of the used queries it is necessary to organize them in a way so it is possible to reuse them. One possibility to organize the queries is to cluster them according to the topic they are related to (content-based clustering).