分辨矩阵法约简中医辨证要素初探

目的 应用分辨矩阵法对慢性支气管炎辨证要素进行约简。方法 对100例慢性支气管炎患者的口唇、发热、呼吸、脉、舌、苔、咳、痰8个决策属性进行初步数字化、数据预处理和试探性挖掘，以得到最简规则。结果 在决策属性的约简中，脉、舌、痰共同构成知识的核，口唇、呼吸两个决策属性被约简。结论 应用分辨矩阵法可得到慢性支气管炎辨证分型的最小决策规则，可作为一种中医诊断数字化的新方法。

Preliminary Study on Reduction of Traditional Chinese Medicine Syndrome Differentiation Factors by Discernibility Matrix Method

Objective To reduce the syndrome differentiation factors for chronic bronchitis by discernibility matrix method. Methods One hundred patients with chronic bronchitis were included in the study. Preliminary digitization, data preprocessing, and exploratory data mining were performed on eight decision attributes, i.e., lip, fever, breath, pulse, tongue, coating on the tongue, cough, and phlegm, to obtain the simplest rule. Results Among the decision attributes, pulse, tongue, and phlegm were core ones, but lip and breath were deleted. Conclusion The minimal decision rule for syndrome differentiation of chronic bronchitis can be acquired by discernibility matrix method, which can be used as a novel method for the digitization of traditional Chinese medicine diagnosis.