Re-ranking search results using an additional retrieved list

Lior Meister, Oren Kurland, Inna Gelfer Kalmanovich
2010 Information retrieval (Boston)  
We present a novel approach to re-ranking a list that was retrieved in response to a query so as to improve precision at the very top ranks. The approach is based on utilizing a second list that was retrieved by using, for example, a different retrieval method and/or query representation. In contrast to commonly-used methods for fusion of retrieved lists, our approach also exploits inter-document-similarities between the lists -a potentially rich source of additional information. Empirical
more » ... ation shows that our methods are effective in re-ranking a high quality TREC run using a second high-quality run; the resultant performance also favorably compares with that of a state-of-the-art fusion method. Furthermore, we show that our methods can help to effectively tackle two long-standing challenges; namely, integration of document-based and cluster-based retrieved results; and, improvement of the performance robustness, and overall effectiveness, of pseudo-feedback-based retrieval.
doi:10.1007/s10791-010-9150-8 fatcat:f3tpdowtofge7cnwtcuvvh4nsa