A physics-guided CNN-BiLSTM model using 15 engineered features achieves RMSE of 19.53 W/m² for GHI forecasting in Sudan, outperforming self-attention models at 30.64 W/m².
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Outperforming Self-Attention Mechanisms in Solar Irradiance Forecasting via Physics-Guided Neural Networks
A physics-guided CNN-BiLSTM model using 15 engineered features achieves RMSE of 19.53 W/m² for GHI forecasting in Sudan, outperforming self-attention models at 30.64 W/m².