Table 3.

Two- and multiclass classifier results.

Analysis type	Two-class analysis			Multiclass analysis
	I	II	III	IV	V
STN area	ZI vs. VMR	ZI vs. DLR	DLR vs. VMR	ZI vs. DLR vs. VMR	ZI vs. DLR vs. VMR
Model	Linear SVM	Linear SVM	Linear SVM	Linear SVM One-vs.-Rest	Linear SVM One-vs.-One
Number of features	25	25	25	25	25
C parameter	0.01	0.1	0.01	0.5	0.5
Precision training set	1 (±0.00)	1 (±0.00)	0.95 (±0.06)	1 (±0.00)	1 (±0.00)
Recall training set	1 (±0.00)	1 (±0.00)	0.88 (±0.08)	1 (±0.00)	1 (±0.00)
Accuracy training set	1 (±0.00)	1 (±0.00)	0.94 (±0.05)	1 (±0.00)	1 (±0.00)
Precision CV test set	1 (±0.00)	0.9 (±0.2)	0.9 (±0.1)	0.94 (±0.1)	0.93 (±0.2)
Recall CV test set	1 (±0.00)	0.875 (±0.3)	0.833 (±0.2)	0.91 (±0.1)	0.92 (±0.2)
Accuracy CV test set	0.98 (±0.05)	0.82 (±0.2)	0.77 (±0.2)	0.82 (±0.3)	0.81 (±0.2)