基于决策树的冠心病痰瘀互阻证识别模式研究

Recognition patterns study of coronary heart disease patients with phlegm-blood stasis syndrome based on decision tree

目的:建立不稳定性心绞痛(UA)患者临床常规检测指标对痰瘀互阻证的识别模式。方法:将临床收集的411例UA患者的基本资料、中医四诊信息及临床常规检测指标进行归一化处理后,采用卡方自动交互检测(CHAID)决策树方法从89个临床常规检测指标中自动提取痰瘀互阻证的识别规律,形成识别模式。结果:高敏C反应蛋白(hsCRP)、总胆红素(TBIL)、谷氨酰转肽酶(GGT)、血小板(PLT)、空腹血糖(Glu)和P-R间期经筛选进入决策树模型。该模型对411例患者的测试结果显示:灵敏度为72.46%,特异度为81.29%,检验正确率为79.81%。收益及指数图表结果显示出该模型的良好性能。结论:决策树模型可以基于临床常规检测指标数据清晰、直观的进行UA患者痰瘀互阻证的识别,自动归纳识别规律,在证型-理化指标对应模式的数据挖掘中具备一定的优势。

Objective： To construct a recognition pattern of UA patients with phlegm-blood stasis based on clinical physicochemical index. Methods： Clinical basic data, four diagnostic information and physicochemical index of 411 UA patients were collected and normalized. CHAID decision tree was applied to self-extract recognition rules from 89 clinical indexes, and recognition pattern of phlegm-blood stasis syndrome was established. Results： hs-CRP, TBIL, GGT, PLT, Glu and P-R interval were selected to enter the pattern. The sensitivity and specificity of the decision tree pattern could reach 72.46% and 81.29%, and the accuracy was 79.81%. Gain and index charts showed the goog performance of the pattern. Conclusion： Decision tree pattern could identify phlegm-blood stasis syndrome of UA patients clearly and more intuitively, and it also could self-extract recognition rules. It had advantage in the data mining of syndrome-clinical physicochemical index corresponding pattern.