Using morphology in English-Turkish statistical machine translation

In this study, statistical machine translation approaches for English-Turkish language couple are offered using morphological analysis in order to achieve different sub-lexical representations on SIU corpus. For language pairs having different inflectional morphologies like English-Turkish language pair, statistical machine translation efforts which use word forms as basic translation unit generally shows low performance. Hence, using different sub-lexical representations instead of word forms improves the machine translation performance.