An initiative of Bangla-English natural language translation using case

Abstract

In this research we figured out a framework for Bangla-English Translator using case. Bangla is an Irregular Language. Head-Driven Phrase Structure Grammar (HPSG) based analysis is observed in the research of The Center for Research on Bangla Language Processing (CRBLP). Phrase structure analysis is very complicated for this kind of language. Moreover, knowledge for understanding the meaning of sentences can be divided into two classes, linguistic and non-linguistic. The representation of non-linguistic knowledge is not easy task for irregular and ambiguous language like Bangla. So, we propose a case structure analysis on verb and propose a case identification mythology which includes three properties; case marker, category and location in the sentence. To perform this, we used some well known linguistic analysis. And for translation into English, we used Phrase structure analysis in SVO (Subject, Verb, Object) form. Classifying the noun words in our proposed categories, an appropriate case can be identified. So, for Bangla-English translation first we are going to make the part-of-speech tagging for the source language then we will perform the case analysis then case frame mapping for source to destination language. After that using a dictionary we will translate the case frames into destination sentence.