DCU@TRECMed 2012: Using ad-hoc baselines for domain-specific retrieval
Leveling, Johannes and Goeuriot, Lorraine and Kelly, Liadh and Jones, Gareth J.F. (2012) DCU@TRECMed 2012: Using ad-hoc baselines for domain-specific retrieval. In: Medical Records Track, Text Retrieval Conference (TREC) 2012, 6-9 Nov 2012, Gaithersburg, USA.
Full text available as:
This paper describes the first participation of DCU in the TREC Medical Records Track (TRECMed). We performed some initial experiments on the 2011 TRECMed data based on the BM25 retrieval model. Surprisingly, we found that the standard BM25 model with default parameters, performs comparable to the best automatic runs submitted to TRECMed 2011 and would have resulted in rank four out of 29 participating groups. We expected that some form of domain adaptation would increase performance. However, results on the 2011 data proved otherwise: concept-based query expansion decreased performance, and filtering and reranking by term proximity also decreased performance slightly. We submitted four runs based on the BM25 retrieval model to TRECMed 2012 using standard BM25, standard query expansion, result filtering, and concept-based query expansion. Official results for 2012 confirm that domain-specific knowledge does not increase performance compared to the BM25 baseline as applied by us.
Archive Staff Only: edit this record