Table 1 Comparison between classical and granule-based forecasting approaches.
From: Electric-load forecasting using interval models based on granularity and justifiable principles
Aspect | Classical forecasting (point-based) | Proposed granule-based forecasting |
|---|---|---|
Forecast output | Single prediction \(\hat{y}_i\) | Interval \([\hat{y}_i - \delta _i, \hat{y}_i + \delta _i]\) |
Uncertainty quantification | Absent | Explicit and interpretable |
Distribution assumption | Often assumed (e.g. Gaussian) | Not required |
Robustness | Sensitive to noise | Robust to variability |
Interpretability | Low | High (semantic meaning of width) |
Metrics | RMSE, MAE | Justification Score |
