moha/AraCovid19-MFH
收藏Hugging Face2025-10-06 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/moha/AraCovid19-MFH
下载链接
链接失效反馈官方服务：
资源简介：
ARACOVID19-MFH是一个包含10828条阿拉伯语推文的手动注释多标签数据集，用于假新闻和仇恨言论检测。这些推文被注释了10个不同的标签，标签设计考虑了推文的可查证性、积极/消极性和事实性等因素。数据集除了用于假新闻检测，还可以应用于仇恨言论检测、观点/新闻分类、方言识别等任务。
ARACOVID19-MFH is a manually annotated multi-label dataset containing 10,828 Arabic tweets for fake news and hate speech detection. The tweets are annotated with 10 different labels, designed to consider aspects such as the tweets checkability, positivity/negativity, and factuality. The dataset is primarily designed for fake news detection but can also be used for hate speech detection, opinion/news classification, dialect identification, and other tasks.
提供机构：
moha
