Solar Power Forecasting on GEFCom Zone 1 2014 (test)
0.994Pearson Correlation CoefficientProposed Method
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Proposed MethodType=Conditional Diffusion + Inpainting, Input Representation=2D Patch + Masked, Probabilistic=✓2026.05 | 0.994 | 1.7 | 0.022 | |
| Diffusion (Unconditional)Type=Generative, Input Representation=Time-Series, Probabilistic=✓2026.05 | 0.99 | 2 | 0.03 | |
| FEDformerType=Transformer, Input Representation=Frequency-aware, Probabilistic=✗2026.05 | 0.989 | 2.1 | 0.032 | |
| AutoformerType=Transformer, Input Representation=Decomposition-based, Probabilistic=✗2026.05 | 0.988 | 2.2 | 0.034 | |
| InformerType=Transformer, Input Representation=Long Sequence, Probabilistic=✗2026.05 | 0.987 | 2.4 | 0.036 | |
| CNN-LSTMType=Hybrid DL, Input Representation=Spatial–Temporal, Probabilistic=✗2026.05 | 0.983 | 2.8 | 0.041 | |
| GRUType=Deep Learning (RNN), Input Representation=Sequential, Probabilistic=✗2026.05 | 0.98 | 3.1 | 0.045 | |
| LSTMType=Deep Learning (RNN), Input Representation=Sequential, Probabilistic=✗2026.05 | 0.978 | 3.4 | 0.048 | |
| SVRType=Machine Learning, Input Representation=Handcrafted Features, Probabilistic=✗2026.05 | 0.962 | 5.2 | 0.071 | |
| ARIMAType=Statistical, Input Representation=1D Time Series, Probabilistic=✗2026.05 | 0.95 | 6.5 | 0.085 |