Details

Autor(en) / Beteiligte

• Singh, Thoudam Doren

Titel

Building Parallel Corpora for SMT System: A Case Study of English-Manipuri

Ist Teil von

• International journal of computer applications, 2012-01, Vol.52 (14), p.47-51

Ort / Verlag

New York: Foundation of Computer Science

Erscheinungsjahr

2012

Quelle

EZB Electronic Journals Library

Beschreibungen/Notizen

• The Statistical Machine Translation (SMT) systems are developed using sentence aligned parallel corpus. The difficulty is that there is no parallel corpus at the required measure for many language pairs. The preparation of large scale parallel corpus takes time and demands the linguistics skill. In the present work, the various issues of a quality parallel corpus and a technique that extracts parallel corpus between Manipuri, a morphologically rich and resource constrained Indian language and English has been developed from a web based comparable news corpora. We explore the crux of the parallel corpora towards improving the translation quality through linguistics factors for the language pair.