Selective Prediction on Diabetic Retinopathy (DR) (test)
0.65AUSCSale_EU_crit
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Sale_EU_critFamily=Per-class var., Model=Bayesian EfficientNet-B4, Number of Bayesian samples (S)=30, Bootstrap resamples=2002026.02 | 0.65 | 40.9 | 84.3 | 0 | |
| MaxProbFamily=Scalar, Model=Bayesian EfficientNet-B4, Number of Bayesian samples (S)=30, Bootstrap resamples=2002026.02 | 0.639 | 43.9 | 84.6 | 0 | |
| Var_critFamily=Per-class var., Model=Bayesian EfficientNet-B4, Number of Bayesian samples (S)=30, Bootstrap resamples=2002026.02 | 0.606 | 37.9 | 84.2 | 0 | |
| EntropyFamily=Scalar, Model=Bayesian EfficientNet-B4, Number of Bayesian samples (S)=30, Bootstrap resamples=2002026.02 | 0.604 | 40.1 | 84.2 | 0 | |
| Sale_EU_globalFamily=Per-class var., Model=Bayesian EfficientNet-B4, Number of Bayesian samples (S)=30, Bootstrap resamples=2002026.02 | 0.457 | 34.1 | 84 | 0 | |
| OvA_MIFamily=Per-class MI, Model=Bayesian EfficientNet-B4, Number of Bayesian samples (S)=30, Bootstrap resamples=2002026.02 | 0.452 | 36.7 | 83.8 | 0 | |
| MIFamily=Scalar, Model=Bayesian EfficientNet-B4, Number of Bayesian samples (S)=30, Bootstrap resamples=2002026.02 | 0.436 | 33.9 | 83.8 | 0 | |
| CBECFamily=Boundary, Model=Bayesian EfficientNet-B4, Number of Bayesian samples (S)=30, Bootstrap resamples=2002026.02 | 0.416 | 33.5 | 84 | 0 | |
| Ccrit_sumFamily=Per-class Ck, Model=Bayesian EfficientNet-B4, Number of Bayesian samples (S)=30, Bootstrap resamples=2002026.02 | 0.327 | 32.1 | 83.1 | 0 | |
| Ccrit_maxFamily=Per-class Ck, Model=Bayesian EfficientNet-B4, Number of Bayesian samples (S)=30, Bootstrap resamples=2002026.02 | 0.285 | 30.2 | 82.7 | 50.5 |