MedChain
收藏资源简介:
MedChain是由香港城市大学、香港中文大学、深圳大学、阳明交通大学和台北荣民总医院联合创建的临床决策数据集,包含12,163个临床案例,涵盖19个医学专科和156个子类别。数据集通过五个关键阶段模拟临床工作流程,强调个性化、互动性和顺序性。数据来源于中国医疗网站“iiYi”,经过专业医生验证和去识别化处理,确保数据质量和患者隐私。MedChain旨在评估大型语言模型在真实临床场景中的诊断能力,解决现有基准在个性化医疗、互动咨询和顺序决策方面的不足。
MedChain is a clinical decision dataset jointly developed by City University of Hong Kong, The Chinese University of Hong Kong, Shenzhen University, National Yang Ming Chiao Tung University, and Taipei Veterans General Hospital. It contains 12,163 clinical cases covering 19 medical specialties and 156 subcategories. The dataset simulates the clinical workflow through five key stages, emphasizing personalization, interactivity and sequentiality. The data is sourced from the Chinese medical website "iiYi", and has been validated by professional physicians and de-identified to ensure data quality and patient privacy. MedChain aims to evaluate the diagnostic capabilities of large language models (LLMs) in real-world clinical scenarios, addressing the limitations of existing benchmarks in personalized healthcare, interactive consultation and sequential decision-making.




