Detecting Abnormal Brain Regions in Schizophrenia Using Structural MRI via Machine Learning

Table 2

Classification performance of different methods based on GM.

Method	ACC	SN	SP	GM	DM	F2M
SVM	0.6277	0.6923	0.5711	0.6288	0.6167	0.6339
2T + SVM	0.6829	0.7226	0.6431	0.6817	0.6741	0.6812
RFE + SVM	0.7326	0.8427	0.6533	0.7420	0.7321	0.7699
PCA + SVM	0.7057	0.7712	0.6521	0.7092	0.7029	0.7103
ICA + SVM	0.6806	0.7114	0.6862	0.6987	0.6571	0.6968
TBFS + SVM	0.7118	0.8200	0.6267	0.7169	0.7155	0.7327
2T + PCA + SVM	0.7981	0.8193	0.7704	0.7945	0.7896	0.7922
2T + ICA + SVM	0.7504	0.7727	0.7111	0.7413	0.7359	0.7394
2T + TBFS + SVM	0.7794	0.8009	0.7484	0.7742	0.7658	0.7701
Ours	0.7962	0.8142	0.7777	0.7957	0.7851	0.7896