Information Retrieval and Trainable Natural Language Processing
暂无分享,去创建一个
The initial goal was to use simple techniques borrowed from the MITRE's message understanding system to improve a standard IR system. The AA. decided to modify SMART behavior via pre- and postprocesses. The preprocess intended to improve Smart is fairly simple. Its aim is to remove material from the queries that is not useful, and is in fact possibily misleading. Another preprocess is used to identified particularly salient keywords and phrases used in each query, based on a number of simple heuristics, beginning with either the title field or the first sentence of the narrative. The postprocess is employed to improve the ordering of the returned documents, based on the presence or absence of the keywords and phrases identified
[1] G. T. Nicol,et al. Flex : the lexical scanner generator , 1993 .
[2] David D. Palmer,et al. Transformation-Based Bracketing: Fast Algorithms And Experimental Results , 1996 .
[3] Chris Buckley,et al. New Retrieval Approaches Using SMART: TREC 4 , 1995, TREC.
[4] Lynette Hirschman,et al. MITRE: Description of the Alembic System Used for MUC-6 , 1995, MUC.
[5] David S. Day,et al. Finite-state phrase parsing by rule sequences , 1996, COLING.