| Autor: |
Mishra, Keshav, Shaikh, Awais, Chauhan, Jyoti, Kanojia, Mahendra |
| Předmět: |
|
| Zdroj: |
International Journal of Computer Information Systems & Industrial Management Applications; 2023, Vol. 15, p227-235, 9p |
| Abstrakt: |
Sanskrit is an ancient language with a rich literary and cultural heritage, but it is not widely spoken today. However, its importance in understanding ancient Indian texts and culture has driven researchers to develop machine translation systems for Sanskrit to English. The goal of these systems is to automatically translate Sanskrit text into English, making it accessible to a wider audience. Language study and the use of human communication languages to interact with machines is a prominent research domain in Natural Language Processing [NLP]. The Sanskrit language being the oldest, we found that there is limited work done to include Sanskrit and its translation using NLP. In this study, we use NLP and Deep learning Transformer based attention mechanisms to translate Sanskrit to English. We have used a corpus dataset to train our model and reported 20% accuracy using the Bhagavad Gita dataset and 72% accuracy using the Bible dataset which can be considered a good standard. As we increase the number of lines in the dataset the Model gives better accuracy. We compared the Transformer Model and Long Short-Term Memory (LSTM) Model. Our model performs better than our previous models used to translate the Sanskrit language. They will also aid the linguistic community in saving the time-consuming process of Sanskrit to English translation. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
