Open Access
American Research Journal of Computer Science and Information Technology
ISSN (Online): 2572-2921
DOI: 10.46568/arjcsit
Sentiment Analysis of Marathi Tweets: A Comparative Study of Transformer Based Models
1Department of Computer Science Engineering, Indian Institute of InformationTechnology, Dharwad.
2Department of Computer Science and Engineering, Graphic Era University.
Vedant Jadhav, Shantanu Tripathi, et al., “Sentiment Analysis of Marathi Tweets: A Comparative Study of
Transformer Based Models”, American Research Journal of Computer Science and Information Technology, Vol 5, no. 1,
2022, pp. 1-5.
Abstract
The Internet is widely utilized as a platform for exchanging information and ideas. In these encounters, a large volume
of textual data is produced. NLP requires a large amount of data for training, which can be found on social networking
platforms such as Facebook, Twitter etc.(Shetty) Sarcasm identification, fake news detection, sentiment analysis, and
other similar tasks have become possible and valuable. We have trained different transformer-based models on Marathi,
a low-resource regional language, and compare their performance in this study (Magueresse et al., 2020). We present
that MuRIL gives better performance than other Transformer based models.
