Research Ideas and Methods on Mathematical Relationship of Syndrome Differentiation in Traditional Chinese Medicine
GAO Wulin
DAI Guohua
JI Yandi
GUAN Hui
DONG Xueyan
Abstract：The mathematical model of syndrome differentiation is the core of intelligent auxiliary diagnosis system of traditional Chinese medicine(TCM).In recent years,the intelligent research of TCM syndrome differentiation with knowledge engineering,mathematical statistics and machine learning as the main methods has been carried out successively.However,due to the multidimensional nonlinearity of TCM syndrome differentiation,there are respective deficiencies in the accuracy and interpretability of the models.The fundamental reason lies in the fact that the mathematical relationship of syndrome differentiation in traditional Chinese medicine is still unclear.This article explores the research ideas and methods of the mathematical relationship in TCM syndrome differentiation by focusing on the mapping relationship between"syndrome and symptoms"and the derivation relation-ship of the"four diagnostic methods".Firstly,Bayesian networks,additive models and decision trees are introduced to clarify the diagnostic value of"symptom-syndrome"in TCM syndrome differentiation,the synergistic effect of"syndrome+syndrome"and the differential characteristics of"syndrome-syndrome".Secondly,by using knowledge graphs,representation learning reasoning combined with the Delphi method,the reasoning path and diagnostic threshold of the"four diagnostic methods"in TCM syndrome differentiation are analyzed to clarify the mathematical relationship of TCM syndrome differentiation,providing a scientific basis for the construction of the mathematical model of TCM syndrome differentiation.
Keywords：traditional Chinese medicine syndrome differentiationmathematical relationshipmapping relationshipthe four diagnostic methodsknowledge graphrepresentation learning reasoning
Publication Date：2025-11-05
Online Publishing Date：2026-05-22(First online date of this platform, not the publication date of the document)
Pages：4( 1196-1199 )