BaTEClaCor: A Novel Dataset for Bangla Text Error Classification and Correction
Published in Association for Computational Linguistics, 2023
Our paper is based upon a project utilizing advanced machine learning models and transformer models such as BanglaBERT and BanglaT5, to detect and correct errors in online Bangla communication. Our rigorously analyzed dataset of 10,000 authentic comments showcases BanglaBERT’s 79.1% accuracy in error classification and BanglaT5’s superior Rouge-L score (0.8459), contributing significantly to linguistic precision in the Bangla-speaking digital community.
Recommended citation: Your Name, You. (2023). "BaTEClaCor: A Novel Dataset for Bangla Text Error Classification and Correction." Association for Computational Linguistics.
Download Paper | Download Slides