Genomic data sampling and its effect on classification performance assessment

BMC Bioinformatics

Table 2 Summary of cross-validation results for SRBCT data: Mean ± standard error of classification accuracies

Train/test runs	50%–50% cross validation	75%–25% cross validation	95%–5% cross validation
10	0.768 ± 0.0169	0.755 ± 0.0169	0.750 ± 0.0745
25	0.747 ± 0.0112	0.749 ± 0.0134	0.740 ± 0.0444
50	0.737 ± 0.0076	0.759 ± 0.0102	0.745 ± 0.0281
100	0.734 ± 0.0059	0.750 ± 0.0079	0.733 ± 0.0202
500	0.733 ± 0.0031	0.749 ± 0.0039	0.747 ± 0.0097
1000	0.732 ± 0.0021	0.752 ± 0.0027	0.749 ± 0.0068
2000	0.734 ± 0.0014	0.748 ± 0.0019	0.751 ± 0.0047
3000	0.733 ± 0.0012	0.749 ± 0.0016	0.758 ± 0.0038
4000	0.733 ± 0.0010	0.747 ± 0.0014	0.760 ± 0.0033
5000	0.733 ± 0.0009	0.746 ± 0.0012	0.761 ± 0.0030

ISSN: 1471-2105