Biomarker Extraction Based on Subspace Learning for the Prediction of Mild Cognitive Impairment Conversion

Table 4

Classification performance of learning projection matrices using different data.

Data of learning projection matrices	Classifier	ACC (%)	value	SEN (%)	SPE (%)	AUC
pMCI and sMCI data	Decision tree classifier	60.89	<0.0001	67.22	52.27	0.5318
AD and NC data		66.35		70.97	60.33	0.6216
pMCI and sMCI data	SVM classifier with RBF kernel	65.17	<0.0001	73.95	53.46	0.6495
AD and NC data		68.37		83.57	48.06	0.6800
pMCI and sMCI data	SVM classifier with linear kernel	64.78	<0.0001	67.59	60.91	0.6628
AD and NC data		69.37		75.39	61.23	0.6951