An enhancement of binary particle swarm optimization for gene selection in classifying cancer classes

Table 5 Comparative Experimental Results of the best subsets produced by EPSO and BPSO

Data	Method	EPSO			BPSO
	Evaluation	Best	#Ave	S.D	Best	#Ave	S.D
11_Tumors	#Acc (%)	96.55	95.40	0.61	97.7	96.78	0.73
	#Genes	243	237.70	9.66	230	221.30	14.48
	#Time	67.21	67.05	0.22	212.44	252.05	32.87
9_Tumors	#Acc (%)	76.67	75.00	1.11	85	82.67	2.25
	#Genes	251	247.10	9.65	250	244.30	9.43
	#Time	4.35	4.29	0.09	17.97	17.74	0.10
Brain_Tumor1	#Acc (%)	93.33	92.11	0.82	93.33	92.44	0.47
	#Genes	8	7.5	2.51	217	215	7.24
	#Time	14.89	14.82	0.12	24.62	24.76	0.15
Brain_Tumor2	#Acc (%)	94	92.4	1.27	92	91.40	0.97
	#Genes	4	6.0	1.83	208	225.20	13.13
	#Time	0.85	0.85	0.002	6	6.08	0.09
Leukemia1	#Acc (%)	100	100	0	100	98.75	0.44
	#Genes	2	3.20	0.63	213.00	198.20	6.48
	#Time	2.80	2.79	0.01	6.97	6.99	0.05
Leukemia2	#Acc (%)	100	100	0	100	100	0
	#Genes	4	6.8	2.2	195	200.80	4.24
	#Time	3.1	3.1	0.02	10.05	10.16	0.15
Lung_Cancer	#Acc (%)	96.06	95.67	0.31	98.03	97.73	0.25
	#Genes	7	8.3	2.11	212	212.90	9.26
	#Time	123.95	124.08	0.49	145.98	145.93	1.00
SRBCT	#Acc (%)	100	99.64	0.58	100	100	0
	#Genes	7	14.90	13.03	191	195.30	3.23
	#Time	7.51	7.51	0.03	22.75	22.63	0.22
Prostate_Tumor	#Acc (%)	99.02	97.84	0.62	98.04	98.04	0.00
	#Genes	5	6.6	2.17	195	199.40	3.41
	#Time	4.37	4.37	0.008	22.57	22.27	0.16
DLBCL	#Acc (%)	100	100	0	100	100	0
DLBCL	#Genes	3	4.7	0.82	193	200.90	3.78
	#Time	1.99	1.99	0.008	7.72	7.75	0.10

Note: The best result of each evaluation is written in a bold style. The best method of each data set is shown in the shaded cells. The best method is selected based on the highest number of the best results for all evaluations. #Acc and S.D. denote the classification accuracy and the standard deviation, respectively, whereas #Genes and #Ave represent the number of selected genes and an average, respectively. #Time stands for running time in the hour unit.

ISSN: 1748-7188