Multi-instance Learning Classification on TIGER classical MIL (10-fold cross-val)
96.1AccuracyTR-RGMIL
Evaluation results
| Method | Links | |
|---|---|---|
| TR-RGMILEvaluation Protocol=10-fold CV, Runs=5 runs2026.05 | 96.1 | |
| ConveyanceEvaluation Protocol=10-fold CV, Runs=5 runs, Design=default parameters and no MIL-specific design choices2026.05 | 90.4 | |
| DPMILEvaluation Protocol=10-fold CV, Runs=5 runs2026.05 | 89.7 | |
| PSMILEvaluation Protocol=10-fold CV, Runs=5 runs2026.05 | 88.4 | |
| DSMILEvaluation Protocol=10-fold CV, Runs=5 runs2026.05 | 86.9 | |
| BDRMILEvaluation Protocol=10-fold CV, Runs=5 runs2026.05 | 86.9 | |
| Gated-ABMILEvaluation Protocol=10-fold CV, Runs=5 runs2026.05 | 84.5 | |
| RGMILEvaluation Protocol=10-fold CV, Runs=5 runs2026.05 | 84.2 | |
| ABMILEvaluation Protocol=10-fold CV, Runs=5 runs2026.05 | 83.9 |