Query Reformulation in Collaborative Information Retrieval

Information retrieval (IR) systems utilize user feedback for generating optimal queries with respect to a particular information need. However the methods that have been developed in IR for generating these queries do not memorize information gathered from previous search processes, and hence can not use such information in new search processes. Thus each new search process does not know anything about previous search processes and can not profit from the results of the previous processes. We call systems which can consider results from previous search processes Collaborative Information Retrieval (CIR) systems. Improving retrieval quality in a CIR system should be possible, since the system can learn from many queries issued from various users. In this paper we present a new method for use in CIR. We are proposing to use previously learned queries and their relevant documents for improving overall retrieval quality. Based on the similarity of a new query to previously learned queries we are expanding the new query by extracting terms from documents which have been judged as relevant to these previously learned queries. Thus our method uses global feedback information for query expansion in contrast to local feedback information which has been used in previous work in query expansion methods.