About this document
Bengali Hate Speech Dataset Overview by Sadman Shifat Protic is a document available to read on EtoBox.
This paper presents a new dataset of 30,000 Bengali comments for hate speech detection, addressing the lack of resources in this area. The dataset includes comments from various categories and was annotated by 50 individuals, achieving a 91.05% accuracy in labeling hate speech. Baseline evaluations showed that while deep learning models performed well, an SVM model achieved the highest accuracy of 87.5%.
- Author
- Sadman Shifat Protic
- Language
- EN