Improving the quality of MT output using novel name entity translation scheme

Name Entity Translation has become a challenge for the machine translators as it has become a cardinal part of Natural Language Processing Applications. Name Entity comprises of two subtasks i.e. they can either be translated or transliterated with the help of syllabification. This paper describes the translation and transliteration of name entities from English to Punjabi using statistical rule based approach. Various rules are constructed with the help of syllabification approach. We are transliterating the name entities by applying the syllabification algorithm. Name entities involved in our experiment are: Proper name, Location name, Organization name and miscellaneous. Transliteration of name entities is obtained with the help of Probability calculation. We have calculated N-Gram probabilities for all the syllables on the basis of relative frequency. For the purpose of probability calculation we have used a statistical machine translation toolkit MOSES.