Skip to main content
. 2007 Dec 18;1(Suppl 1):S56. doi: 10.1186/1753-6561-1-s1-s56

Table 3.

Paired Wilcoxon rank test of prediction errors from three RFs, using Replicates 1–100

Training data Test data


Comparison of prediction errors p-Value Difference in median p-Value Difference in median
ITbp vs. ITtop50 3.94 × 10-18 -1.40 9.09 × 10-10 0.43
ITbp vs. IT0 3.95 × 10-18 -3.33 2.57 × 10-12 -0.73
ITtop50 vs. IT0 3.94 × 10-18 -1.87 1.20 × 10-17 -2.10