RoCliCo
收藏arXiv2023-10-10 更新2024-06-21 收录
下载链接:
https://github.com/dariabroscoteanu/RoCliCo
下载链接
链接失效反馈官方服务:
资源简介:
RoCliCo是首个公开的罗马尼亚语点击诱饵检测数据集,由布加勒斯特大学计算机科学系创建。该数据集包含8,313个新闻样本,均由人工标注为点击诱饵或非点击诱饵。数据来源于罗马尼亚的六个公共新闻网站,确保了训练和测试数据的不重叠。数据集的创建过程涉及详细的标注指南和本地罗马尼亚语使用者的参与。RoCliCo的应用领域主要集中在自动检测误导性新闻标题,以保护在线用户的时间不被浪费,同时也为罗马尼亚语的点击诱饵检测研究提供了重要的资源。
RoCliCo is the first publicly available Romanian clickbait detection dataset, created by the Department of Computer Science of the University of Bucharest. This dataset contains 8,313 news samples, all manually labeled as either clickbait or non-clickbait. The data is sourced from six public Romanian news websites, ensuring non-overlapping training and test datasets. The dataset creation process involved detailed annotation guidelines and the participation of native Romanian speakers. The main application scenarios of RoCliCo focus on automatically detecting misleading news headlines to protect online users from wasting their time, while also providing a valuable resource for Romanian clickbait detection research.
提供机构:
布加勒斯特大学计算机科学系
创建时间:
2023-10-10



