HERDPhobia

HERDPhobia是由尼日利亚巴伊亚大学和HausaNLP创建的，旨在识别针对尼日利亚Fulani族群的仇恨言论的数据集。该数据集包含6174条推文，涵盖英语、尼日利亚皮钦语和豪萨语三种语言。数据收集通过关键词搜索，并按照严格的标注指南分为仇恨、非仇恨和不确定三类。HERDPhobia主要用于自然语言处理领域，特别是仇恨言论的自动检测，以帮助解决社交媒体中的仇恨言论问题。HERDPhobia is a dataset created by Bayero University, Nigeria and HausaNLP, which aims to identify hate speech targeting the Fulani ethnic group in Nigeria. This dataset contains 6,174 tweets covering three languages: English, Nigerian Pidgin, and Hausa. The data was collected via keyword searches, and classified into three categories—hate speech, non-hate speech, and uncertain—according to strict annotation guidelines. HERDPhobia is primarily applied in the field of natural language processing, particularly for automatic hate speech detection, to help address the problem of hate speech on social media.