NTUBROWS System for NTCIR-7 Information Retrieval for Question Answering

This paper presents an information retrieval system for the NTCIR-7 information retrieval for question answering task. This system is composed by three parts: (1) Query processing (2) Retrieval model (3) Re-rank module. Query processing filters stop-words and selects query terms to generate a required term set for further retrieval. Threes retrieval models from two famous retrieval systems are adopted. They retrieved relevant documents based on the generated required set. Finally, Re-rank module gives documents scores according to the distributions of their possessive terms. The performance of our system achieves a mean average precision (MAP) of 0.4635, a Q-measure of 0.4811, and an MSn-DCG of 0.6831 on NTCIR-7 IR4QA testing set.