Pathological Brain Detection Using Weiner Filtering, 2D-Discrete Wavelet Transform, Probabilistic PCA, and Random Subspace Ensemble Classifier

Table 5

Classification comparison (DS-66, DS-160, and DS-255).

Approaches	Feature	Run	Accuracy (%)
DWT + SVM + POLY [24]	4761		DS-66	DS-160	DS-255
DWT + SVM + RBF [24]	4761	5	98.00	97.15	96.37
DWT + PCA + -NN [4]	7	5	98.00	97.33	96.18
DWT + PCA + FNN + ACPSO [32]	19	5	98.00	97.54	96.79
DWT + PCA + FNN + SCABC [33]	19	5	100.00	98.75	97.38
DWT + PCA + BPNN + SCG [7]	19	5	100.00	98.93	97.81
DWT + PCA + KSVM [5]	19	5	100.00	98.29	97.14
RT + PCA + LS-SVM [34]	9	5	100.00	99.38	98.82
SWT + PCA + IABAP-FNN [11]	7	10	100.00	98.88	98.43
WT + PCA + ABC-SPSO-FNN [11]	7	10	100.00	99.44	99.18
WE + NBC [35]	7	10	92.58	99.62	99.02
DWT + PCA + ADBRF [17]	13	5	100.00	99.30	98.44
DWT + SUR + ADBSVM [18]	7	5	100.00	99.22	98.43
FRFE + DP-MLP + ARCBBO [16]	12	10	100.00	99.19	98.24
FRFE + BDP-MLP + ARCBBO [16]	12	10	100.00	99.31	98.12
DWT + PCA + RSE	13	5	100.00	99.57	98.90
DWT + PPCA + RSE (proposed)	13	5	100.00	100.00	99.20