Research on classification of prescriptions in Treatment Record of Hundreds of Selected Ancient Prescriptions based on hierarchical clustering / 国际中医中药杂志

ABSTRACT

Objective:Based on systematic idea of clustering methods, a novel approach for classifying prescriptions was explored in the light of the method of classifying analogous prescriptions.Methods:A total of 581 ancient prescriptions were selected from Treatment Record of Hundreds of Selected Ancient Prescriptions, and the standardized names of Traditional Chinese Medicines in these prescriptions were entered in Microsoft Excel 2007, then imported to SPSS 24.0 to generate dendrograms using the hierarchical clustering function. The classification of 581 selected ancient prescriptions was analyzed. Results:All 581 prescriptions could be classified into 86 categories through repeated clustering, the largest group has 29 prescriptions, the smallest group has 2 prescriptions, and the average group has about 6.75 prescriptions. In general, the later the intercepted group, the lower the similarity of its internal prescriptions.Conclusion:This method could realize the classification of prescriptions in the light of the method of classifying analogous prescriptions, which may help to break the original thinking bondage and further deepen the understanding of compatibility rules of prescriptions. However, its disadvantage lied in that the theoretical clues would be reduced when analyzing the compatibility rules of prescriptions and the issues of drug dosage, nature, taste and meridian categories were not specially considered.