Study of TCM Syndrome Identification Modes for Patients with Type 2 Diabetes Mellitus Based on Data Mining

Table 3

Results of three classification methods for 554 cases with T2-DM syndrome.

Method	T2-DM syndrome	TN	FP	Sensitivity (%)	Specificity (%)	Accuracy (%)	AUC
		FN	TP
LR	SHF syndrome	233	28	89.3	90.1	89.7	0.953
	QYD syndrome	29	264
QUEST algorithm of DT	SHF syndrome	204	57	78.2	91.5	85.2	0.931
	QYD syndrome	25	268
CHAID algorithm of DT	SHF syndrome	204	57	78.2	91.5	85.2	0.931
	QYD syndrome	25	268
KNN	SHF syndrome	211	40	84.7	91.5	88.3	0.887
	QYD syndrome	15	278

Sensitivity = TP/(TP + FN); specificity = TN/(TN + FP); accuracy = (TP + TN)/(TP + FN + TN + FP).