A Novel Approach to Ensemble Classifiers: FsBoost-Based Subspace Method

Table 9

Results for the FsBoost.V2 ensemble algorithm (Levels 1 and 2).

Level	Level 1									Level 2
Classifier	Ensemble 1			Ensemble 2			Ensemble 3			Ensemble
For A dataset
FS	0			1			2			0/1/2
NF	68			25			15			68/25/15
Class	Sen	Spe	Acc	Sen	Spe	Acc	Sen	Spe	Acc	Sen	Spe	Acc
E	0.92	1.00	95.91	0.96	1.00	98.00	0.91	1.00	95.52	0.94	1.00	97.22
N-E	1.00	0.92		1.00	0.96		1.00	0.91		1.00	0.94
AUC	0.96			0.98			0.96			0.97
Kappa	0.92			0.96			0.91			0.94
F-measure	0.96			0.98			0.95			0.97
For B dataset
FS	0			1			2			0/1/2
NF	68			25			15			68/25/15
Class	Sen	Spe	Acc	Sen	Spe	Acc	Sen	Spe	Acc	Sen	Spe	Acc
E	0.90	1.00	94.83	0.89	0.99	94.17	0.83	0.99	91.09	0.88	1.00	94.04
N-E	1.00	0.90		0.99	0.89		0.99	0.83		1.00	0.88
AUC	0.95			0.94			0.91			0.94
Kappa	0.90			0.88			0.82			0.88
F-measure	0.95			0.94			0.90			0.94
For C dataset
FS	0			1			2			0/1/2
NF	68			25			15			68/25/15
Class	Sen	Spe	Acc	Sen	Spe	Acc	Sen	Spe	Acc	Sen	Spe	Acc
E	0.87	1.00	93.30	0.88	0.99	93.39	0.90	0.96	93.35	0.90	0.99	94.52
N-E	1.00	0.87		0.99	0.88		0.96	0.90		0.99	0.90
AUC	0.93			0.93			0.93			0.95
Kappa	0.87			0.87			0.87			0.89
F-measure	0.93			0.93			0.93			0.94
For D dataset
FS	0			1			2			0/1/2
NF	68			25			15			68/25/15
Class	Sen	Spe	Acc	Sen	Spe	Acc	Sen	Spe	Acc	Sen	Spe	Acc
E	0.84	0.97	90.39	0.85	0.97	91.09	0.86	0.94	89.57	0.87	0.97	91.65
N-E	0.97	0.84		0.97	0.85		0.94	0.86		0.97	0.87
AUC	0.90			0.91			0.90			0.92
Kappa	0.81			0.82			0.79			0.83
F-measure	0.90			0.91			0.89			0.91

Sen: sensitivity, Spe: specificity, Acc: accuracy (%), FS: feature selection, NF: number of features, EC: ensemble classifier, E: epilepsy, and N-E: nonepilepsy.