Table 2.

The classification performance for all prediction models analyzed in this study

model	dataset	AUC	ACC (%)	SPE (%)	SEN (%)	NPV (%)	PPV (%)
7-miRPairs	Training	0.987	98.47	98.14	99.25	99.67	95.82
	Test	0.974	97.22	96.87	98.02	99.14	93.03
139-miRPairs	Training	1.000	98.95	98.58	99.81	99.92	96.80
	Test	0.998	98.14	97.59	99.44	99.75	94.62
106-miRPairs	Training	1.000	98.90	98.50	99.81	99.92	96.63
	Test	0.997	97.97	97.47	99.15	99.63	94.35
253-miRPairs	training	1.000	100	100	100	100	100
	Test	0.986	98.23	97.71	99.44	99.75	94.88
5-miRNAs	Training	0.779	82.53	89.47	66.38	86.09	73.06
	Test	0.782	82.70	89.41	66.95	86.40	72.92

AUC area under the curve, ACC accuracy, SPE specificity, SEN sensitivity, NPV negative predictive value, PPV positive predictive value