جستجوی مقالات مرتبط با کلیدواژه « ترجمه خودکار » در نشریات گروه « ادبیات و زبان ها »

Today, linguistic corpora play a crucial role in solving different types of linguistic issues thanks to emerging computer technology. Bilingual parallel corpora aligned at sentence and word level can be retrieved for single-word as well as multi-word units making easier further applications in different computer and language areas. In this paper we address the problem of exploiting English-Persian parallel corpus in making an efficient bilingual concordance using mutual information measure. Here, a mutual information statistics is used to add word level alignments between English and Persian sentence pairs in our parallel corpus. A parallel corpus with alignments on the word level has certainly many applications among which phrase-based translation memory software, terminology management, cross-language information retrieval, statistical machine translation system and the like. We conducted an experiment using our algorithm and compared alignment outputs with manually aligned sentences. Experimental results revealed that our concordancing program gained the accuracy rate of 75% which seems very encouraging.