The Hong Kong Polytechnic University at the TREC 2004 Robust Track

In the robust track, we mainly tested our passage-based retrieval model with different passage sizes and weighting schemes. In our approach, we used two retrieval models, namely the 2-Poisson model using BM25 term weights and the vector space model (VSM) using adaptive pivoted unique document length normalization. Also, we utilize WordNet to re-weight some PRF terms and extract some context words as expanded query terms. We show that our passage-based model achieves the comparable performance on the whole query set. Moreover, our new methods of using WordNet information for query expansion can improve the retrieval performance.

• Type: Conference
• Year: 2004
• Where: TREC
• Authors: D. Y. Wang, Robert Wing Pong Luk, Kam-Fai Wong
• Keywords: Passage-based Retrieval Model, Retrieval Models, TREC 2004, TREC 2008, Vector Space Model