Binary Classification on Income (test)
87.39Test AccuracyNon-Private
Evaluation results
| Method | Links | |
|---|---|---|
| Non-Privateprivacy_setting=Non-Private2026.01 | 87.39 | |
| PAC-private responsesper-step MI budget (b)=infinity2026.01 | 87.17 | |
| PAC-private responsesper-step MI budget (b)=2^-42026.01 | 87.15 | |
| PAC-private responsesper-step MI budget (b)=2^-82026.01 | 86.68 | |
| lightGBMSensitive Attribute=race, Evaluation Protocol=5-fold cross-validation2023.01 | 86.1 | |
| lightGBMSensitive Attribute=sex, Evaluation Protocol=5-fold cross-validation2023.01 | 86.1 | |
| FairGBMSensitive Attribute=sex, Evaluation Protocol=5-fold cross-validation2023.01 | 86.08 | |
| FairGBMSensitive Attribute=race, Evaluation Protocol=5-fold cross-validation2023.01 | 86.02 | |
| PAC-private responsesper-step MI budget (b)=2^-122026.01 | 85.92 | |
| PAC-private responsesper-step MI budget (b)=2^-162026.01 | 85.86 | |
| Retrainm (number of unlearning classes)=2, n (percentage of samples to unlearn)=25%2025.12 | 85.84 | |
| PAC-private responsesper-step MI budget (b)=2^-202026.01 | 85.84 | |
| PAC-private responsesper-step MI budget (b)=2^-242026.01 | 85.84 | |
| PAC-private responsesper-step MI budget (b)=2^-282026.01 | 85.84 | |
| PAC-private responsesper-step MI budget (b)=2^-322026.01 | 85.84 | |
| FedORAm (number of unlearning classes)=2, n (percentage of samples to unlearn)=25%2025.12 | 85.48 | |
| BaggingSensitive Attribute=race, Evaluation Protocol=5-fold cross-validation2023.01 | 84.15 | |
| BaggingSensitive Attribute=sex, Evaluation Protocol=5-fold cross-validation2023.01 | 84.15 | |
| POAFSensitive Attribute=race, Evaluation Protocol=5-fold cross-validation2023.01 | 83.8 | |
| POAFSensitive Attribute=sex, Evaluation Protocol=5-fold cross-validation2023.01 | 83.8 | |
| EPAF-DSensitive Attribute=race, Evaluation Protocol=5-fold cross-validation2023.01 | 83.55 | |
| EPAF-DSensitive Attribute=sex, Evaluation Protocol=5-fold cross-validation2023.01 | 83.55 | |
| EPAF-CSensitive Attribute=race, Evaluation Protocol=5-fold cross-validation2023.01 | 83.53 | |
| EPAF-CSensitive Attribute=sex, Evaluation Protocol=5-fold cross-validation2023.01 | 83.53 | |
| Retrainm (number of unlearning classes)=1, n (percentage of samples to unlearn)=50%2025.12 | 82.91 | |
| CVFUm (number of unlearning classes)=1, n (percentage of samples to unlearn)=50%2025.12 | 82.84 | |
| FedORAm (number of unlearning classes)=1, n (percentage of samples to unlearn)=50%2025.12 | 82.6 | |
| CVFUm (number of unlearning classes)=2, n (percentage of samples to unlearn)=25%2025.12 | 81.18 | |
| ICOm (number of unlearning classes)=2, n (percentage of samples to unlearn)=25%2025.12 | 80.98 | |
| ICOm (number of unlearning classes)=1, n (percentage of samples to unlearn)=50%2025.12 | 80.92 | |
| GAm (number of unlearning classes)=1, n (percentage of samples to unlearn)=50%2025.12 | 79.36 | |
| GAm (number of unlearning classes)=2, n (percentage of samples to unlearn)=25%2025.12 | 78.91 | |
| AdaFairSensitive Attribute=race, Evaluation Protocol=5-fold cross-validation2023.01 | 24.89 | |
| AdaFairSensitive Attribute=sex, Evaluation Protocol=5-fold cross-validation2023.01 | 24.89 |