BanglaAbuseMeme

BanglaAbuseMeme数据集由印度理工学院卡拉格普尔分校创建，旨在支持自动检测孟加拉语辱骂性表情包。该数据集包含4043个孟加拉语表情包，其中1515个被标记为辱骂性，其余为非辱骂性。每个表情包都附有详细标签，包括是否粗俗、是否讽刺、情感倾向（积极、中立、消极）以及目标社区。数据集通过关键词搜索和社交媒体平台收集，旨在帮助研究人员和开发者构建和评估检测辱骂性内容的模型，以提高网络环境的安全性。

The BanglaAbuseMeme dataset was developed by the Indian Institute of Technology Kharagpur to support automatic detection of Bengali abusive memes. It includes 4043 Bengali memes, of which 1515 are annotated as abusive and the remaining are categorized as non-abusive. Each meme is paired with detailed annotations covering whether it contains obscenity, whether it is sarcastic, its sentiment polarity (positive, neutral, negative), and the targeted communities. The dataset was collected via keyword searches and social media platforms, and aims to assist researchers and developers in constructing and evaluating abusive content detection models, thereby enhancing the safety of online environments.