Binary Classification (AD vs CN) on sMRI
82.93AccuracyAgent Ensemble
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Agent EnsembleArchitecture=NeuroAgent Ensemble, Ensemble Mode=stratified stacking of modality-specific predictions, Validation=5-fold cross-validation2026.05 | 82.93 | 77.86 | 65.11 | 70.92 | 86.24 | 0.5944 | |
| ResNet-18Architecture=3D CNN, Validation=5-fold cross-validation2026.05 | 73.33 | 58.55 | 56.81 | 57.67 | 78.15 | 0.3822 | |
| EfficientNet-B0Architecture=3D CNN, Validation=5-fold cross-validation2026.05 | 73.13 | 57.28 | 62.77 | 59.9 | 77.35 | 0.3985 | |
| DenseNet-121Architecture=3D CNN, Validation=5-fold cross-validation2026.05 | 53.74 | 39.77 | 86.81 | 54.55 | 62.93 | 0.254 |