Table 4 Statistical comparison of evaluation metrics
Hypothesis | AUROC (p-value) | |
|---|---|---|
Null | Alternative | |
PM = LR | PM > LR | < 0.05 (PM has a significantly better AUROC than LR) |
PM = LR | LR > PM | > 0.05 (LR does not have a significantly better AUROC than PM) |
PM = RF | PM > RF | < 0.05 (PM has a significantly better AUROC than RF) |
PM = RF | RF > PM | > 0.05 (RF does not have a significantly better AUROC than PM) |
RF = LR | RF > LR | > 0.05 (RF does not have a significantly better AUROC than LR) |
RF = LR | LR > RF | > 0.05 (LR does not have a significantly better AUROC than RF) |