Using mutual information technique in cross-language information retrieval

Abstract

This paper describes the Indonesian-English cross language information system, where we investigated the problem of cross language document retrieval for Indonesian-English. Our work based on Indonesian-English parallel corpus and applied mutual information technique. Our parallel corpus was built using a machine translation tool. Indonesian queries were translated into English by means of bilingual word list that created using mutual information technique. Experimental results demonstrated that translating Indonesian queries into English using this bilingual word list achieved 41.9% of the monolingual queries. © 2008 Springer Berlin Heidelberg.