English to Marathi Machine Translation Linguistic Divergence Using ANN

Abstract

The field of machine translation is included in Natural Language Procession. The development of MT system Marathi to English language is much more challenging task because of linguistic divergences. The most significant aspect of MT is that it maximizes meaning, such that understanding the output requires the least amount of effort and time. In machine translation different Types of Technique Available but as per Literature review Neural Machine Translation gives the Best results. In This paper focus on three types of divergences and also Artificial Neural Network techniqueMachine translation's major problem of divergence has led to the development of neural machine translation (NMT), a different strategy to address the same problemIn this study, used a neural machine translation technique to construct a separate module. Here create thematic divergence module, Categorical divergence module and structural divergence module with 66.67 %, 71.43 %, and 51.25 % accuracy and Calculated BLEU score.