Computational Representation of Paninian Rules of Sanskrit Grammar for Dictionary-Independent Machine Translation

Abstract

Since the beginning of computational linguistics machine translation is being one of the holy grail of natural language processing. With the advancement of world wide web and emergence of data driven methods the machine translation has received a jolt of new activity and more visibility. Still the visible lack of accuracy in machine translation shows that the much work remains to be done in concern of human identical translation. The approaches of machine translation uses the dictionary meaning of the target language as a heart of translation process. Dictionary based machine translation system continuously requires the intervention from the human expert. Unlike other languages Sanskrit is having a standalone procedure for developing new words and defining its meaning as well. Maharshi Panini developed a complete set of grammatical rules that defines the whole procedure to process the input language. The paper presents the computer simulation of Paninian rules of Sanskrit grammar focuses on the generation of meaning of unknown word using most elementary component of a language - dhAtu. We have used the direct machine translation approach for achieving dictionary-independent machine translation.