Table 3:

Accuracy metrics for predicting disease status

Disease	Model	Sensitivity	Specificity	Balanced accuracy	Misclassification	PPV1	NPV2
Healthy	iRF125	87.2	81.9	84.6	16.2	30.6	53.1
HCC		81.8	87.2	84.5	14.4	24.3	61.3
Cirrhosis		33.3	92.9	63.1	21.6	8.1	70.3
Average		67.4	87.3	77.4	17.4	21	61.6
Healthy	iRF12	87.2	85	86.1	14.1	34.3	51.5
HCC		84.8	92.4	88.6	10.1	28.3	61.6
Cirrhosis		63	91.7	77.3	16.2	17.2	66.7
Average		78.3	89.7	84	13.5	26.6	59.9
Healthy	iRF4	87.2	93.3	90.3	9.1	34.3	56.6
HCC		87.9	95.4	91.7	7.1	29.3	63.6
Cirrhosis		81.5	90.3	85.9	12.1	22.2	65.7
Average		85.5	93	89.3	9.4	28.6	62
Healthy	Decision Tree	82	80.3	81.2	19.1	29.1	51.8
HCC		87.9	93.5	90.7	8.2	26.4	65.4
Cirrhosis		81.5	90.4	85.9	11.8	20	68.2
Average		83.8	88.1	85.9	13	25.2	61.8

¹PPV=positive predictive value;

²NPV=negative predictive value