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

Table 3 Experimental result for each using run epso on 11_tumors, 9_tumors, brain_tumor1, brain_tumor2 and leukemia1 data sets

Run#	11_Tumors		9_Tumors		Brain_Tumor1		Brain_Tumor2		Leukemia1
	#Acc (%)	#Selected genes	#Acc (%)	#Selected genes	#Acc (%)	#Selected genes	#Acc (%)	#Selected genes	#Acc (%)	#Selected genes
1	96.55	243	76.67	251	93.33	8	94	4	100	3
2	95.98	245	76.67	255	93.33	11	94	5	100	4
3	95.98	250	75	231	92.22	6	94	8	100	3
4	95.40	232	75	237	92.22	7	92	4	100	3
5	95.40	241	75	242	92.22	8	92	7	100	3
6	95.40	244	75	253	92.22	9	92	7	100	4
7	94.83	218	75	255	92.22	11	92	7	100	4
8	94.83	229	75	261	91.11	3	92	7	100	3
9	94.83	232	73.33	238	91.11	5	92	8	100	3
10	94.83	243	73.33	248	91.11	7	90	3	100	2
Average ± S.D.	95.40 ±0.61	237.70 ±9.66	75 ±1.11	247.10 ±9.65	92.11 ±0.82	7.5 ±2.51	92.40 ±1.26	6.00 ±1.83	100.00 ±0	3.20 ±0.63

Note: Results of the best subsets are written in a bold style. A near-optimal subset that produces the highest classification accuracy with the smallest number of genes is selected as the best subset. #Acc and S.D. denote the classification accuracy and the standard deviation, respectively, whereas #Selected Genes and Run# represent the number of selected genes and a run number, respectively.

ISSN: 1748-7188