Table 12 Holm post-hoc test results based on the G-means
From: Boosting methods for multi-class imbalanced data classification: an experimental review
Control method (CatBoost) | ||||
|---|---|---|---|---|
Algorithm | Z | p-value | Holm | Hypothesis (α = 0.05) |
AdaBoost.MH | 2.1385 | 0.0324 | 0.0062 | Not rejected |
SAMME | 1.9857 | 0.047 | 0.0071 | Not rejected |
LogitBoost | 0.6328 | 0.5268 | 0.05 | Not rejected |
GradientBoost | 1.7675 | 0.0771 | 0.01 | Not rejected |
XGBoost | 1.1565 | 0.2474 | 0.0166 | Not rejected |
MEBoost | 2.1385 | 0.0324 | 0.0055 | Not rejected |
SMOTEBoost | 0.7637 | 0.445 | 0.025 | Not rejected |
RUSBoost | 2.5531 | 0.0106 | 0.0038 | Not rejected |
LightGBM | 2.2912 | 0.0219 | 0.005 | Not rejected |
AdaC1 | 2.5313 | 0.0113 | 0.0041 | Not rejected |
AdaC2 | 1.7021 | 0.0887 | 0.0125 | Not rejected |
AdaC3 | 2.3349 | 0.0195 | 0.0045 | Not rejected |
AdaCost | 1.833 | 0.0667 | 0.0083 | Not rejected |