Table 9 LOOCV three-class misclassification errors with the original Minnesota data and the percentiles of LOOCV errors with 50 permuted datasets by SC.
From: A comparative study of discriminating human heart failure etiology using gene expression profiles
# of top genes | Original data | Permutated data | |||||
|---|---|---|---|---|---|---|---|
| Â | LOOCV errors | P-value | 0% | 25% | 50% | 75% | 100% |
50 | 11 | .00 | 13 | 17 | 20 | 22.75 | 29 |
100 | 11 | .00 | 12 | 17.25 | 20 | 22 | 30 |
400 | 14 | .02 | 13 | 18 | 20.5 | 23 | 28 |
1600 | 13 | .00 | 13 | 18 | 20 | 23 | 29 |
6400 | 13 | .02 | 12 | 18 | 20 | 23 | 29 |