资源 · 2026
TCMKG: Knowledge-Grounded question answering in Traditional Chinese Medicine via dynamic prompt evolution framework
Journal of King Saud University - Computer and Information Sciences
「TCMKG: Knowledge-Grounded question answering in Traditional Chinese Medicine via dynamic prompt evolution framework」是 中医大模型开源资源目录（Awesome-TCM-LLM）收录的资源，记录于 2026。该条目属于中医大模型（TCM LLM）领域，与 GitHub 列表同源维护。
TCMKG在新安医案上用专家引导提示抽取实体，建成含37192个实体、130030条关系的中医知识图谱，信息抽取精确率0.89、召回率0.88、F1为0.88。