Phrasal Queries with LingPipe and Lucene: Ad Hoc Genomics Text Retrieval.
Bob Carpenter · Text REtrieval Conference · 2004
The hypothesis we explored for the Ad Hoc task of the Genomics track for TREC 2004 was that phrase-level queries would increase precision over a baseline of token-level terms. We implemented our approach using two open source tools: the Apache Jakarta Lucene TF/IDF search engine (version 1.3) and the Alias-i LingPipe tokenizer and namedentity annotator (version 1.0.6). Contrary to our intuitions, the baseline system provided better performance in terms of recall and precision for almost every query at almost every precision/recall operating point.