Bengali Toxic Comments Dataset

该数据集包含16,073个孟加拉语评论实例，这些评论被手动标注为有毒和无毒两类。有毒评论进一步细分为六个子类别：粗俗、仇恨、宗教、威胁、恶意挑衅和侮辱。该数据集由专业标注员手动分类，确保了高质量的数据标注，对于有毒和无毒类别，其平均kappa分数达到了0.96。规模上，数据集包含了16,073个实例，任务是对有毒评论进行多标签分类。

This dataset contains 16,073 Bengali review instances, which are manually annotated into two categories: toxic and non-toxic. Toxic reviews are further subdivided into six subcategories: vulgar, hate speech, religious, threatening, malicious provocation, and insulting. Manually classified by professional annotators, the dataset ensures high-quality data annotation, with an average Cohen's Kappa score of 0.96 for both the toxic and non-toxic categories. With a total of 16,073 instances, the task of this dataset is multi-label classification for toxic comments.