NELA-GT-2018
收藏arXiv2019-04-03 更新2024-06-21 收录
下载链接：
https://doi.org/10.7910/DVN/ULHLCB
下载链接
链接失效反馈官方服务：
资源简介：
NELA-GT-2018数据集由丹麦技术大学创建，包含713,534篇文章，收集自2018年2月至11月，涵盖194个新闻和媒体来源。该数据集独立于社交媒体互动，通过8个不同的评估站点整合了来源级别的真实性评级，包括可靠性、偏见、透明度和消费者信任等多个维度。此数据集旨在解决广泛标记数据集的缺乏问题，支持机器学习和混合方法研究，以全面理解误导性和极端主义新闻生产者的影响和策略。
The NELA-GT-2018 dataset was created by the Technical University of Denmark, containing 713,534 articles collected between February and November 2018 across 194 news and media outlets. This dataset is independent of social media interactions, and integrates source-level authenticity ratings from 8 distinct evaluation platforms, covering multiple dimensions such as reliability, bias, transparency, and consumer trust. It aims to address the scarcity of broadly labeled datasets, and supports machine learning and mixed-methods research to comprehensively understand the impacts and strategies of misinformation and extremist news producers.
提供机构：
丹麦技术大学
创建时间：
2019-04-03
