A Framework on Performance Analysis of Mathematical Model-Based Classifiers in Detection of Epileptic Seizure from EEG Signals with Efficient Feature Selection

Table 5

Average MSE for Haar, dB4, and Sym 8 wavelet features in various classifiers with PSO feature selection.

Wavelets	Classifiers	TP	TN	FP	FN	MSE
Haar	Linear regression	52	61	39	48	0.000212
	Nonlinear regression	64	55	45	36	0.00011
	Gaussian mixture model (GMM)	51	74	26	49	0.000231
	KNN	55	87	13	45	8.67E − 05
	SVM (linear)	56	53	47	44	0.0002
	SVM (polynomial)	82	80	20	18	1.23E − 05
	SVM (RBF)	85	89	11	15	3.48E − 06
dB4	Linear regression	71	81	19	29	2.12E − 05
	Nonlinear regression	72	77	23	28	2.23E − 05
	Gaussian mixture model (GMM)	55	78	22	45	0.000106
	KNN	87	82	18	13	7.44E − 06
	SVM (linear)	66	76	24	34	3.21E − 05
	SVM (polynomial)	63	66	34	37	4.9E − 05
	SVM (RBF)	63	93	7	37	1.96E − 05
Sym8	Linear regression	59	89	11	41	3.64E − 05
	Nonlinear regression	57	52	48	43	0.000233
	Gaussian mixture model (GMM)	64	51	49	36	0.00029
	KNN	54	52	48	46	0.000336
	SVM (linear)	53	56	44	47	0.000191
	SVM (polynomial)	55	59	41	45	0.000126
	SVM (RBF)	90	90	10	10	1.96E − 06