Monant Medical Misinformation Dataset

Monant Medical Misinformation Dataset专注于医疗新闻文章和博客，由Kempelen Institute of Intelligent Technologies创建，旨在通过机器学习方法解决COVID-19时代医疗错误信息的激增问题。该数据集包含约317,000篇医疗新闻文章/博客和3,500个经过事实核查的声明，以及573个手动和超过51,000个自动标记的声明与文章之间的映射。数据集不仅支持声明存在检测和文章立场分类等任务，还适用于错误信息特征研究、错误信息传播分析和来源可靠性分类等。通过提供多种模态数据和详细的元数据，该数据集为多模态方法的研究提供了可能，并有助于理解错误信息如何在不同语言和来源间传播。

The Monant Medical Misinformation Dataset, developed by the Kempelen Institute of Intelligent Technologies, focuses on medical news articles and blogs, and was created to address the surge of medical misinformation during the COVID-19 pandemic via machine learning approaches. This dataset comprises approximately 317,000 medical news articles/blogs, 3,500 fact-checked claims, as well as 573 manually labeled and over 51,000 automatically labeled mappings between claims and their corresponding articles. It supports a range of tasks including claim presence detection and article stance classification, and is also suitable for research on misinformation characteristics, misinformation propagation analysis, and source reliability classification, among other related tasks. By offering multimodal data and detailed metadata, this dataset enables research on multimodal methodologies and facilitates insights into how misinformation propagates across diverse languages and source types.