Knowledge Mining from Clinical Datasets Using Rough Sets and Backpropagation Neural Network

Table 12

Performance measure of best reducts of each dataset.

Dataset	Reduct	TP	FN	TN	FP	Accuracy (%)	Sensitivity (%)	Specificity (%)	TPR	FPR	AUC
Hepatitis	Reduct-	26	2	117	2	97.30	98.32	97.28	0.93	0.017	0.9492
Breast cancer	Reduct-	238	3	451	7	98.60	98.76	98.57	0.99	0.015	0.9952
Heart disease	Reduct-	102	18	142	8	90.40	94.67	90.37	0.85	0.053	0.9204