Word Proximity QA System.
Philip Rennert · Text REtrieval Conference · 2001
This is a question answering system with very little NLP, based on question-category-dependent selection and weighting of search terms, selecting answer strings centered around words most commonly found near search terms. Its performance was medium, with a 0.2 MRR. Certain category strategies may be of interest to other QAers, and are described.