A simple rule-based part of speech tagger
ANLC '92 Proceedings of the third conference on Applied natural language processing
Hi-index | 0.00 |
QA-LaSIE was the heart of the University of Sheffield entry to the Question Answering track of TREC-9. By relaxing some of the strongest linguistic constraints, we achieved a very significant performance improvement over our TREC-8 system on both the TREC-8 and TREC-9 tasks. Whereas most systems returned answers that were always close to the maximum allowable length, our system was one of the only entries that tried to return an "exact answer" to a question.