Effective selection of informative SNPs and classification on the HapMap genotype data

BMC Bioinformatics

Table 3 The maximum classification accuracy in each of 30 simulations together with the mean accuracy (standard deviation), and the relevant feature numbers leading to the maximal accuracy together with the mean number (standard deviation), for 3 populations and 4 populations, respectively

Feature Numbers	Maximum accuracy (%)	Mean accuracy ± std (%)	Relevant feature numbers	Average number of features ± std
f-statistics on 3 populations	94.29 98.57 95.71 97.14 94.29	96.05 ± 1.58 (%)	12 85 15 42 29 86	63.6 ± 25.8
	94.29 95.71 95.71 95.71 97.14		37 56 78 90 71 46
	97.14 95.71 97.14 95.71 98.57		79 79 49 8 75 74
	92.86 97.14 97.14 97.14 92.86		74 100 83 50 81
	97.14 95.71 95.71 95.71 97.14		67 38 82 81 93 84
	95.71 98.57 95.71 97.14 92.86		53
f-statistics on 4 populations	78.57 78.57 85.71 70.00 88.57	77.34 ± 6.57 (%)	53 98 82 96 91	85.2 ± 15.1
	78.57 80.00 82.86 70.00 80.00		100 93 81 99 90
	65.71 78.57 82.86 68.57 81.43		59 99 88 55 73 81
	80.00 70.00 78.57 81.43 74.29		56 99 100 100 74
	72.86 84.26 74.29 84.29 88.57		74 98 94 88 72 81
	68.57 75.71 64.29 75.71 70.00		99 98 90
Modified t-test on 3 populations	95.71 100.00 95.71 98.57	97.09 ± 1.74 (%)	27 84 19 90 29 95	64.0 ± 26.5
	94.29 98.57 95.71 95.71		64 80 84 80 78 54
	95.71 98.57 95.71 95.71 98.57		83 92 57 11 80 79
	100.00 100.00 97.14 98.57		78 95 62 75 94 84
	95.71 97.14 98.57 98.57 95.71		53 10 32 28 68 55
	94.29 94.29 97.14 97.14 95.71
	98.57 97.14 98.57
Modified t-test on 4 populations	82.86 77.14 87.14 82.86 82.86	83.86 ± 3.16 (%)	31 84 92 94 86	84.1 ± 16.3
	84.29 88.57 80.00 81.43 82.86		100 84 99 83 87
	82.86 84.29 87.14 84.29 85.71		96 61 83 93 99 95
	81.43 82.86 84.29 84.29 75.71		80 81 88 99 99 75
	81.43 84.29 81.43 84.29 90.00		95 82 95 51 83 73
	84.29 85.71 85.71 85.71 90.00		99 55

ISSN: 1471-2105