Table 6.

Prediction performance of PCR-based model using five-gene panel with various machine learning approaches (SMO, sequential minimal optimization).

	Naive Bayes (%)	J48 (%)	SMO (%)	Random forest (%)
Accuracy	62	53	36	40
Sensitivity	83	83	89	92
Specificity	61	49	29	31
PPV	23	19	15	16
NPV	96	95	95	97