Skip to main content

Table 7 Pair-wise Wilcoxon signed-ranks test for L5 results (W is the test statistic)

From: Multi-task learning with a natural metric for quantitative structure activity relationship learning

SettingWp-value
STL vs feature-based MTL medians: 0.744 and 0.7013746461.609e−13
STL vs instance-based MTL medians: 0.744 and 0.6335351972.2e−16
Feature-based MTL vs instance-based MTL medians: 0.701 and 0.6335356732.2e−16