Multimodal AD vs CN Classification on Tau-PET + MRI (five folds)
85.65AccuracyAgent Ensemble
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Agent Ensembletype=Stratified stacking, components=Baselines + MLP stacker2026.05 | 85.65 | 82.46 | 70 | 75.72 | 91.17 | 0.6607 | |
| ResNet-18architecture=ResNet-182026.05 | 77.76 | 67.65 | 58.3 | 62.63 | 84.55 | 0.4718 | |
| EfficientNet-B0architecture=EfficientNet-B02026.05 | 76.46 | 63.3 | 62.77 | 63.03 | 80.9 | 0.4577 | |
| DenseNet-121architecture=DenseNet-1212026.05 | 59.46 | 43.45 | 88.94 | 58.38 | 69.45 | 0.3387 |