Table 1 Summary of cross-validation results for leukaemia data: Mean ± standard error of classification accuracies
From: Genomic data sampling and its effect on classification performance assessment
Train/test runs | 50%–50% cross validation | 75%–25% cross validation | 95%–5% cross validation |
|---|---|---|---|
10 | 0.786 ± 0.0176 | 0.844 ± 0.0319 | 0.775 ± 0.0583 |
25 | 0.760 ± 0.0149 | 0.804 ± 0.0225 | 0.730 ± 0.0477 |
50 | 0.732 ± 0.0117 | 0.763 ± 0.0164 | 0.735 ± 0.0306 |
100 | 0.725 ± 0.0083 | 0.744 ± 0.0125 | 0.750 ± 0.0213 |
500 | 0.736 ± 0.0033 | 0.764 ± 0.0049 | 0.770 ± 0.0093 |
1000 | 0.740 ± 0.0025 | 0.766 ± 0.0034 | 0.770 ± 0.0066 |
2000 | 0.739 ± 0.0018 | 0.764 ± 0.0024 | 0.768 ± 0.0048 |
3000 | 0.738 ± 0.0015 | 0.762 ± 0.0019 | 0.767 ± 0.0038 |
4000 | 0.740 ± 0.0013 | 0.762 ± 0.0017 | 0.770 ± 0.0033 |
5000 | 0.741 ± 0.0011 | 0.762 ± 0.0015 | 0.771 ± 0.0029 |