Calibration Error Estimation on Synthetic binary distribution predictor f (test)
0.016Mean Estimated Calibration ErrorSCDL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SCDLalpha (Distribution parameter)=0, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.016 | 0.003 | |
| smCEalpha (Distribution parameter)=0, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.021 | 0.014 | |
| smCEalpha (Distribution parameter)=1, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.025 | 0.016 | |
| smCEalpha (Distribution parameter)=0.8, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.027 | 0.016 | |
| smCEalpha (Distribution parameter)=0.5, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.028 | 0.013 | |
| Cutoffalpha (Distribution parameter)=0, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.03 | 0.012 | |
| SCDLalpha (Distribution parameter)=0.5, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.036 | 0.006 | |
| bECEalpha (Distribution parameter)=0, N (Number of samples)=1000, n (Training samples)=500, Number of bins=112026.05 | 0.043 | 0.011 | |
| bECEalpha (Distribution parameter)=1, N (Number of samples)=1000, n (Training samples)=500, Number of bins=112026.05 | 0.064 | 0.065 | |
| Cutoffalpha (Distribution parameter)=0.5, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.068 | 0.016 | |
| SCDLalpha (Distribution parameter)=1, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.076 | 0.034 | |
| SCDLalpha (Distribution parameter)=0.8, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.08 | 0.014 | |
| Cutoffalpha (Distribution parameter)=0.8, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.11 | 0.016 | |
| bECEalpha (Distribution parameter)=0.5, N (Number of samples)=1000, n (Training samples)=500, Number of bins=112026.05 | 0.117 | 0.015 | |
| Cutoffalpha (Distribution parameter)=1, N (Number of samples)=1000, n (Training samples)=5002026.05 | 0.136 | 0.015 | |
| bECEalpha (Distribution parameter)=0.8, N (Number of samples)=1000, n (Training samples)=500, Number of bins=112026.05 | 0.14 | 0.054 |