Positive-Unlabeled Classification on 20News (test)
83.34AccuracyPU ET
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PU ETArchitecture=Tree-based, Risk estimator=nonnegative, Loss function=quadratic2022.10 | 83.34 | — | |
| nnPUArchitecture=Neural Network2022.10 | 74.11 | — | |
| Self-PUArchitecture=Neural Network, Self-calibration=true2022.10 | 62.87 | — | |
| uPUArchitecture=Neural Network2022.10 | 44.95 | — | |
| nnPUArchitecture=Neural Network2022.10 | — | 78.05 | |
| PU ETRisk Estimator=nonnegative, Loss Function=quadratic2022.10 | — | 85.33 | |
| Self-PUArchitecture=Neural Network, self-calibration=true2022.10 | — | 71.78 | |
| uPUArchitecture=Neural Network2022.10 | — | 12.95 |