A Possibilistic Query Translation Approach for Cross-Language Information Retrieval

In this paper, we explore several statistical methods to find solutions to the problem of query translation ambiguity. Indeed, we propose and compare a new possibilistic approach for query translation derived from a probabilistic one, by applying a classical probability-possibility transformation of probability distributions, which introduces a certain tolerance in the selection of word translations. Finally, the best words are selected based on a similarity measure. The experiments are performed on CLEF-2003 French-English CLIR collection, which allowed us to test the effectiveness of the possibilistic approach