FACTOID

FACTOID数据集由德国马尔堡大学的CAISA实验室创建，专注于分析Reddit上的政治讨论，特别是2020年初至今的虚假新闻传播者。该数据集包含超过4,150名用户和340万条Reddit帖子，不仅包括用户的二元标签，还详细记录了用户的可信度水平（从极低到极高）和政治偏见强度（从极右到极左）。这是首个同时捕捉用户历史帖子的长期上下文及其互动的虚假新闻传播者数据集。数据集的应用领域包括识别虚假信息传播者，通过分析用户的社会联系和心理语言特征，以及研究政治偏见对信息传播的影响，旨在解决社交媒体上的虚假信息问题。

The FACTOID dataset was developed by the CAISA Lab at the University of Marburg, Germany, with a focus on analyzing political discussions on Reddit, especially disinformation spreaders since early 2020. This dataset encompasses more than 4,150 users and 3.4 million Reddit posts. It not only provides binary labels for users but also records detailed information on their credibility levels (ranging from extremely low to extremely high) and the intensity of their political biases (spanning from far-right to far-left). As the first dataset of disinformation spreaders that simultaneously captures the long-term context of users’ historical posts and their social interactions, the FACTOID dataset can be applied to identify disinformation spreaders via analyzing users’ social connections and psycholinguistic features, as well as to investigate the impact of political bias on information dissemination, ultimately aiming to tackle the issue of disinformation on social media platforms.