以《医方类聚·消渴门》为例探讨知识元标引在方证和药证规律研究中的应用

  目的  通过挖掘《医方类聚·消渴门》中的方证和药证规律, 探寻知识元标引在中医古籍数字化中的应用思路和方法。  方法  采用知识元标引方法将《医方类聚·消渴门》进行深度标引, 利用数据库读取标引数据, 将导出的方药逻辑数据进行优化加工, 通过构建知识图谱, 对方证、药证关联进行可视化分析及展示。  结果  《医方类聚·消渴门》共有616个知识体、3 213个知识元、10 298个语义以及4 369个语义关联。方-证知识图谱提示《医方类聚·消渴门》中消渴的主要证候是口渴多饮、小便频多、咽干舌燥、五心烦热, 次要证候是形体消瘦、心神烦躁、小便白浊、小便不利、多食善饥、小便黄赤等。文章以核心证候为例, 通过其与对应的方药社团进行进一步重叠筛选, 获得了治疗“口渴多饮和心神烦躁”的方剂社团组合:清心莲子饮、大麦汤、地骨皮饮、黄连猪肚丸、青粱米饮方等; 中药社团组合节点度高的有:天花粉、甘草、麦冬、黄连、人参、知母、黄芪和石膏。  结论  基于知识元标引的知识图谱分析能层次分明地展示方-证-药之间的关联, 能直观地发现核心证候与治疗所用方药之间的关系，对中医古籍的数字化具有重要的应用价值。 

Exploring the Application of Knowledge Element Indexing in the Study of the Laws of Formula-Syndrome and Medicine-Syndrome with the Example of Categorized Collection of Medical Formulas: Emaciation-Thirst Disease

  OBJECTIVE  To explored the application of knowledge element indexing in the digitization of Chinese medical books by probing into the laws of formula-syndrome and medicine-syndrome in the Categorized collection of medical formulas: Emaciation-thirst disease.  METHODS  The knowledge element indexing method was used to complete the depth indexing of the Categorized collection of medical formulas: Emaciation-thirst disease. In addition, The indexed data from the database was read. On top of that, the logical data of the derived formula and medicines was optimized, and the associations of the formula-syndrome and medicine-syndrome were visualized and analyzed via the construction of knowledge mapping.  RESULTS  There were 616 knowledge bodies, 3 213 knowledge elements, 10 298 semantic meanings, and 4 369 semantic associations in the Categorized collection of medical formulas: Emaciation-thirst disease. The formula-syndrome knowledge mapping suggested that the main syndromes recorded in that chapter included thirsty with strong desire to drink, frequent urination, dry throat and tongue, as well as feverish sensations in palms, soles and chest, while the minor syndromes were emaciation, restlessness in the mind, white turbid urine, dysuria, excessive food intaking with rapid hungering, and dark urine. The study took the main syndromes as an example, and further overlapped and filtered them with the corresponding formula-medicine associations to obtain combination of formula associations for the treatment of "thirsty with strong desire to drink and restlessness in the mind". The combination of formula associations contained Qingxin Lianzi decoction, Damai decoction, Digupi beverage, Huanglian Zhudu pill, and Qingliangmi beverage. Besides, the nodes with high degree in medicine associations included snakegourd root, licorice root, dwarf lilyturf tuber, coptis rhizome, ginseng, common anemarrhena rhizome, astragalus root, and gypsum.  CONCLUSION  The knowledge mapping analysis based on knowledge element indexing can not only show the association between formula-syndrome-medicine in a clear hierarchy, but also directly reveal the relationship between the core syndromes and the formulas used for treatment. Thus, the study has vital application value for digitization of Chinese medical books.