bilawalriaz/MedAtoms-Wiki

MedAtoms是一个开放的原子医学事实数据集，专为构建可靠的医疗AI系统而设计。该子集包含从医学和健康相关维基百科文章中提取的高置信度原子事实，广泛覆盖医学主题并使用易于理解的语言。数据集包含74,376个事实，分为训练集(59,500)、验证集(7,437)和测试集(7,439)。每个事实条目包含唯一ID、文本内容、字数统计、字符数统计、医学领域、标签等信息。数据集覆盖23个医学领域和9,167个独特主题标签，适用于健康信息系统、患者医疗聊天机器人、医学教育等多种应用场景。

MedAtoms is an open dataset of atomic medical facts designed for building reliable healthcare AI systems. This subset contains high-confidence atomic facts extracted from medical and health-related Wikipedia articles, providing broad coverage of medical topics in accessible language. The dataset consists of 74,376 facts, split into training (59,500), validation (7,437), and test (7,439) sets. Each fact entry includes unique ID, text content, word count, character count, medical area, tags, and other metadata. Covering 23 medical areas and 9,167 unique topic tags, the dataset is suitable for various applications including health information systems, patient-facing medical chatbots, and medical education.