Improving the rule based machine translation system using sentence simplification (english to tamil)

The ultimate aim of this research is to develop a Rule Based Machine Translation System (RBMT) using sentence simplification. The sentence pattern for English is SVO and Tamil is SOV. Complex and larger sentence are not easy to parse and translate. So, the sentence simplifier is also accommodated in the rule based system to split a large sentence into simple multiple sentences. Machine translation is the process of translating sentence from one language to another language. Here, English is the source language for the translation system and Tamil is the target language. During the translation process, the system need to learn and get trained by linguistic rules. These rules are classified into two types, namely reordering rules and morphological rules. A bilingual dictionary has been created to support this automatic translation. Linguistic information act as a backbone for the proposed system. To evaluate the performance of the system, we have experimented by testing 250 sentence patterns and we got the overall accuracy about 0.7186. Module wise human evaluation has been done to understand the issues in each modules.