Classification on Diabetes
85.71AccuracyReSS
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ReSSData augmentation=true, Reasoning path structure=x → z → y2026.04 | 85.71 | — | |
| BaseBackbone=XGBoost, n (number of samples)=253k, p (number of features)=212025.03 | 84.9 | — | |
| OpenFEBackbone=XGBoost, n (number of samples)=253k, p (number of features)=212025.03 | 84.9 | — | |
| CAAFEBackbone=XGBoost, n (number of samples)=253k, p (number of features)=212025.03 | 84.9 | — | |
| FeatLLMBackbone=XGBoost, n (number of samples)=253k, p (number of features)=212025.03 | 84.9 | — | |
| OCTreeBackbone=XGBoost, n (number of samples)=253k, p (number of features)=212025.03 | 84.9 | — | |
| LLM-FEBackbone=XGBoost, n (number of samples)=253k, p (number of features)=212025.03 | 84.9 | — | |
| SPNScenario=Non-strategic2026.05 | 82.74 | — | |
| SPNScenario=Strategic2026.05 | 82.41 | — | |
| Direct RLReasoning path structure=x → z → y2026.04 | 82.25 | — | |
| Chunked TabPFNScenario=Non-strategic2026.05 | 81.64 | — | |
| MLPScenario=Non-strategic2026.05 | 81.42 | — | |
| ReSSData augmentation=false, Reasoning path structure=x → z → y2026.04 | 81.39 | — | |
| TabPFN v2.5Scenario=Non-strategic2026.05 | 81.22 | — | |
| TabDPTScenario=Non-strategic2026.05 | 81.19 | — | |
| Drift-Resilient TabPFNScenario=Non-strategic2026.05 | 81.03 | — | |
| LightGBMScenario=Non-strategic2026.05 | 80.96 | — | |
| TabFlexScenario=Non-strategic2026.05 | 80.92 | — | |
| CatBoostScenario=Non-strategic2026.05 | 80.71 | — | |
| TabICLScenario=Non-strategic2026.05 | 80.46 | — | |
| XGBoostScenario=Non-strategic2026.05 | 80.08 | — | |
| SVMScenario=Non-strategic2026.05 | 79.92 | — | |
| Linear modelsScenario=Non-strategic2026.05 | 79.84 | — | |
| XGBoost2026.04 | 78.78 | — | |
| TabICLScenario=Strategic2026.05 | 78.62 | — | |
| TabDPTScenario=Strategic2026.05 | 78.41 | — | |
| TabFlexScenario=Strategic2026.05 | 78.01 | — | |
| MLPScenario=Strategic2026.05 | 77.94 | — | |
| CSVMtype=transductive2025.04 | 77.9 | — | |
| LightGBMScenario=Strategic2026.05 | 77.88 | — | |
| Peernoise_rates=0.1, 0.3, prior_equalization=p != 0.52019.10 | 77.8 | — | |
| Chunked TabPFNScenario=Strategic2026.05 | 77.68 | — | |
| TabNet2026.04 | 77.49 | — | |
| Drift-Resilient TabPFNScenario=Strategic2026.05 | 77.4 | — | |
| DRC + SFTReasoning path structure=x → z → y2026.04 | 77.06 | — | |
| S-KLRSparsity selection protocol=sparsest of 3 most accurate2025.12 | 77 | 52.9 | |
| TabPFN v2.5Scenario=Strategic2026.05 | 76.93 | — | |
| S-KLRSparsity selection protocol=Standard2025.12 | 76.7 | 72.4 | |
| SVMSparsity selection protocol=sparsest of 3 most accurate2025.12 | 76.7 | 53.1 | |
| Linear modelsScenario=Strategic2026.05 | 76.63 | — | |
| Decision Tree2026.04 | 76.62 | — | |
| CatBoostScenario=Strategic2026.05 | 76.54 | — | |
| TabPFN2026.04 | 76.19 | — | |
| TSVMtype=transductive, library=scikit-learn2025.04 | 76 | — | |
| SVMScenario=Strategic2026.05 | 75.85 | — | |
| l1/2-KLRSparsity selection protocol=sparsest of 3 most accurate2025.12 | 75.5 | 74.7 | |
| XGBoostScenario=Strategic2026.05 | 75.02 | — | |
| Peernoise_rates=0.1, 0.3, prior_equalization=p = 0.52019.10 | 74.5 | — | |
| SVMSparsity selection protocol=Standard2025.12 | 74.5 | 49 | |
| BQ-SVMKernel=linear, Noise Level=0%2026.04 | 73.9 | — | |
| Direct SFTReasoning path structure=x → y2026.04 | 73.59 | — | |
| EN-SVMKernel=linear, Noise Level=0%2026.04 | 73.5 | — | |
| BAEN-SVMKernel=linear, Noise Level=0%2026.04 | 73.2 | — | |
| ALS-SVMKernel=linear, Noise Level=0%2026.04 | 72.9 | — | |
| BALS-SVMKernel=linear, Noise Level=0%2026.04 | 72.9 | — | |
| ε-BAEN-SVMKernel=linear, Noise Level=0%2026.04 | 72.4 | — | |
| Pin-SVMKernel=linear, Noise Level=0%2026.04 | 70.6 | — | |
| BAEN-SVMKernel=linear, Noise Level=25% label noise2026.04 | 68.4 | — | |
| l1/2-KLRSparsity selection protocol=Standard2025.12 | 67.8 | 17.1 | |
| BQ-SVMKernel=linear, Noise Level=25% label noise2026.04 | 67.4 | — | |
| BALS-SVMKernel=linear, Noise Level=25% label noise2026.04 | 67 | — | |
| ε-BAEN-SVMKernel=linear, Noise Level=25% feature noise2026.04 | 66.9 | — | |
| EN-SVMKernel=linear, Noise Level=25% label noise2026.04 | 66.5 | — | |
| BALS-SVMKernel=RBF, Noise level=25%, Noise type=label noise2026.04 | 66.5 | — | |
| BAEN-SVMKernel=RBF, Noise level=25%, Noise type=label noise2026.04 | 66.5 | — | |
| ε-BAEN SVMKernel=RBF, Noise level=25%, Noise type=label noise2026.04 | 66.5 | — | |
| ε-BAEN-SVMKernel=linear, Noise Level=25% label noise2026.04 | 66.3 | — | |
| ALS-SVMKernel=RBF, Noise level=25%, Noise type=label noise2026.04 | 66.3 | — | |
| EN-SVMKernel=RBF, Noise level=25%, Noise type=label noise2026.04 | 66.3 | — | |
| Pin-SVMKernel=RBF, Noise level=25%, Noise type=label noise2026.04 | 66.2 | — | |
| ALS-SVMKernel=linear, Noise Level=25% label noise2026.04 | 66.1 | — | |
| BQ-SVMKernel=RBF, Noise level=25%, Noise type=label noise2026.04 | 65.8 | — | |
| EN-SVMKernel=linear, Noise Level=25% feature noise2026.04 | 65.5 | — | |
| IVMSparsity selection protocol=Standard2025.12 | 65.3 | 0.2 | |
| ALS-SVMKernel=linear, Noise Level=25% feature noise2026.04 | 65.3 | — | |
| BQ-SVMKernel=linear, Noise Level=25% feature noise2026.04 | 65.2 | — | |
| BALS-SVMKernel=linear, Noise Level=25% feature noise2026.04 | 65 | — | |
| BAEN-SVMKernel=linear, Noise Level=25% feature noise2026.04 | 64.9 | — | |
| Pin-SVMKernel=linear, Noise Level=25% feature noise2026.04 | 64.8 | — | |
| Pin-SVMKernel=linear, Noise Level=25% label noise2026.04 | 64.4 | — |