Query Expansion for Cross Language Information Retrieval Improvement
Benoît Gaillard, Jean-Léon Bouraoui, Emilie Guimier de Neef, Malek Boualem
This paper is devoted to a new method that uses query expansion to improve multilingual information retrieval. The backbone is an Information Retrieval (IR) system based on a search engine and a multilingual module based on statistical machine translation of metadata. To this system is added a Query Expansion (QE) module which mainly uses linguistic resources to perform the expansion. The aim is to use QE to overcome the limitations of machine translation, and to retrieve more relevant results. The authors demonstrate, with examples, the usefulness of such a system. They also validate it with several measures, which show a clear reduction of the silence for results. This work is part of the Orange Labs VSE project and the QUAERO/MSSE project for which a Multilingual Multimedia Information Retrieval (MMIR) prototype has been designed.
Cross-Language Information Retrieval, Machine Translation, Query Expansion, Natural Language Processing.