Table 5. Classification performance by preprocessing conditions and algorithms.

AUC: area under the curve; RF: random forest; MLP: multi-layer perceptron; AdaBoostM1: adaptive boosting

Preprocess	Algorithm	Accuracy	Sensitivity	Specificity	κ coefficient	AUC
Standardization	J48	0.915	0.178	0.138	0.195	0.637
	RF	0.918	0.134	0.138	0.264	0.877
	MLP	0.920	0.065	0.225	0.204	0.936
	AdaBoostM1	0.909	0.083	0.100	0.153	0.866
	Mean	0.915	0.115	0.150	0.204	0.829
Standardization + synthetic minority oversampling technique	J48	0.913	0.876	0.918	0.819	0.927
	RF	0.916	0.857	0.927	0.824	0.973
	MLP	0.906	0.884	0.916	0.807	0.970
	AdaBoostM1	0.921	0.898	0.916	0.833	0.964
	Mean	0.914	0.879	0.919	0.821	0.959