Table 2.
One-sided Wilcoxon signed rank test results for testing whether the Type-I error distribution of experiments generated under the parameter setting specified in the row has lower mean of ranks compared to the distribution of Type-I errors generated under the parameter setting specified in the column, where the null hypothesis states that there is no difference
| Wilcoxon signed rank test
P-values |
|||||||||
|---|---|---|---|---|---|---|---|---|---|
| wGC | wM | wGCM | woGCM | wGC | wM | wGCM | woGCM | ||
| Non-expressed(EOO,woIF) | Non-expressed(NOOB,woIF) | ||||||||
| wGC | 2.2e-16 | 2.2e-16 | 2.2e-16 | 2.2e-16 | 2.2e-16 | 2.2e-16 | |||
| wM | 2.2e-16 | 2.2e-16 | |||||||
| wGCM | 2.2e-16 | 2.2e-16 | 2.2e-16 | 2.2e-16 | |||||
| woGCM | |||||||||
| Non-expressed(EOO,wIF) | Non-expressed(NOOB,wIF) | ||||||||
| wGC | 1.9e-04 | 2.2e-16 | 2.2e-16 | 1.004e-14 | 2.2e-16 | 6.524e-15 | |||
| wM | 2.2e-16 | 2.2e-16 | 2.2e-16 | ||||||
| wGCM | |||||||||
| woGCM | 1.97e-04 | 2.39e-11 | |||||||
| Expressed(EOO,woIF) | Expressed(NOOB,woIF) | ||||||||
| wGC | 5.47e-12 | 1.2e-12 | |||||||
| wM | 1.18e-09 | 5.5e-12 | 1.75e-09 | 1.2e-12 | |||||
| wGCM | 5.51e-10 | 1.17e-09 | 5.5e-12 | 3.75e-10 | 5.38e-10 | 1.2e-12 | |||
| woGCM | |||||||||
| Expressed(EOO,wIF) | Expressed(NOOB,wIF) | ||||||||
| wGC | 1.43e-04 | 3.93e-03 | |||||||
| wM | 1.14e-09 | 2.78e-06 | 7.88e-10 | 2.57e-09 | 7.80e-06 | 1.75e-09 | |||
| wGCM | 1.15e-09 | 7.70e-10 | 2.56e-09 | 1.75e-09 | |||||
| woGCM | |||||||||
*
Note: A P-value presented in the cell indicates that setting in the corresponding row has a lower mean of ranks in Type-I error distribution than the setting in the corresponding column; if the cell is empty the opposite holds. The P-values are less than or equal to the actual test result. The best parameter setting in each experiment is shown in bold.