Table 5. Comparison of different kernels.

Kernel	Accuracy	Sensitivity	Specificity	AUC
Linear	91.6%	93.1%	90.1%	0.944
Gaussian	91.5%	88.9%	93.1%	0.973
Quadratic	89.4%	88.9%	89.7%	0.952
Cubic	95.7%	94.4%	96.6%	0.980
Cubic with PCA features	93.6%	88.9%	96.6%	0.973