Table 3.

Performance of the 5 selected models with different data normalization methods.

Classifier	Not normalized		Z-score		Min-Max		L2-norm
	Training	Test	Training	Test	Training	Test	Training	Test
Logistic regression	0.823	0.817	0.810	0.777	0.809	0.780	0.840	0.840
Support vector machine (linear)	0.816	0.823	0.819	0.803	0.801	0.807	0.833	0.840
Support vector machine (radial basis function)	0.833	0.823	0.807	0.793	0.829	0.787	0.834	0.833
Random forest	0.796	0.803	0.796	0.803	0.796	0.803	0.789	0.830
Extreme gradient boosting	0.821	0.840	0.821	0.840	0.837	0.837	0.803	0.837