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

Table 2

Classification results of two competing biomarkers using different classifiers.

Feature	Classifier	ACC (%)	value	SEN (%)	SPE (%)	AUC
Global grading biomarker in [15]	Decision tree classifier	64.72	0.0082	69.30	58.56	0.6097
Proposed biomarkers		66.35		70.97	60.33	0.6216
Global grading biomarker in [15]	SVM classifier with RBF kernel	68.22	0.5169	83.35	48.03	0.6790
Proposed biomarkers		68.37		83.57	48.06	0.6800
Global grading biomarker in [15]	SVM classifier with linear kernel	67.43	<0.0001	70.04	63.96	0.6981
Proposed biomarkers		69.37		75.39	61.23	0.6951