Constructing Misinformation Datasets

本数据集由南加州大学创建，专注于从社交媒体话语中构建大规模的错误信息标记数据集。数据集包含约14600条数据，主要来源于Twitter，涵盖了COVID-19疫苗相关的内容。创建过程中，研究者利用新闻源可信度标签作为弱标签，并通过模型引导的标签精炼方法来构建数据集。该数据集旨在解决社交媒体上错误信息的快速识别和分类问题，支持相关领域的研究和政策制定。This dataset was created by the University of Southern California, focusing on constructing a large-scale misinformation-annotated dataset from social media discourse. It contains approximately 14,600 entries primarily sourced from Twitter, covering content related to COVID-19 vaccines. During the dataset construction process, researchers utilized news source credibility labels as weak labels and adopted model-guided label refinement methods to build the dataset. This dataset aims to address the rapid identification and classification of misinformation on social media, supporting research and policy-making in relevant fields.