Dynamic Selection of Information Retrieval Algorithms

Users often search in information retrieval systems iteratively. We claim that changing the underlying retrieval algorithm and query expansion algorithm during these iterations yields improved results. In the rst part of this paper, results of preliminary experiments are presented which indicate the viability of our claim. In the second part, we focus on our approach for the dynamic selection of the proper retrieval algorithms and query expansion algorithms. Our approach is based on case-based planning: A retrieval algorithm sequence which holds certain quality criteria is set up due to the consideration of previous experiences and cases.