Paper: Document Expansion Based on WordNet for Robust IR

ACL ID C10-2002
Title Document Expansion Based on WordNet for Robust IR
Venue International Conference on Computational Linguistics
Session Poster Session
Year 2010

The use of semantic information to im- prove IR is a long-standing goal. This pa- per presents a novel Document Expansion method based on a WordNet-based system to find related concepts and words. Ex- pansion words are indexed separately, and when combined with the regular index, they improve the results in three datasets over a state-of-the-art IR engine. Consid- ering that many IR systems are not robust in the sense that they need careful fine- tuning and optimization of their parame- ters, we explored some parameter settings. The results show that our method is spe- cially effective for realistic, non-optimal settings, adding robustness to the IR en- gine. We also explored the effect of doc- ument length, and show that our method is specially successful with shorter docu- ments.