Classification on Covertype (Cohen's kappa)
0.79Cohen's KappaJARF
Evaluation results
| Method | Links | |
|---|---|---|
| JARFEvaluation protocol=CV splits2025.12 | 0.79 | |
| SPORFEvaluation protocol=CV splits2025.12 | 0.633 | |
| CCFEvaluation protocol=CV splits2025.12 | 0.631 | |
| XGBEvaluation protocol=CV splits2025.12 | 0.622 | |
| RotFEvaluation protocol=CV splits2025.12 | 0.616 | |
| RFEvaluation protocol=CV splits2025.12 | 0.612 | |
| LDA+RFEvaluation protocol=CV splits2025.12 | 0.607 | |
| PCA+RFEvaluation protocol=CV splits2025.12 | 0.602 | |
| TabPFN-MarginLabel budget (N)=100, Batch size=10, Strategy=Margin, Base model=TabPFN2026.03 | 0.471 | |
| TabPFN-RandomLabel budget (N)=100, Batch size=10, Strategy=Random, Base model=TabPFN2026.03 | 0.447 | |
| TabPFN-HybridLabel budget (N)=100, Batch size=10, Strategy=Hybrid, Base model=TabPFN2026.03 | 0.445 | |
| TabPFN-CoresetLabel budget (N)=100, Batch size=10, Strategy=Coreset, Base model=TabPFN2026.03 | 0.439 | |
| CatBoost-MarginLabel budget (N)=100, Batch size=10, Strategy=Margin, Base model=CatBoost2026.03 | 0.42 | |
| XGBoost-MarginLabel budget (N)=100, Batch size=10, Strategy=Margin, Base model=XGBoost2026.03 | 0.388 | |
| TabPFN-Proxy-HybridLabel budget (N)=100, Batch size=10, Strategy=Proxy-Hybrid, Base model=TabPFN2026.03 | 0.381 | |
| LabelSpreading-RandomLabel budget (N)=100, Batch size=10, Strategy=Random, Base model=LabelSpreading2026.03 | 0.24 |