REDf: A Deep Learning Model for Short-Term Load Forecasting to Facilitate Renewable Integration and attaining the SDGs 7, 9, and 13

• Figure 3: Comparison of the PJME dataset before and after normalization. (a) data in its original (without normalization) forms and (b) in normalized (with normalization) form

  

• Figure 6: Interface of the model deployment application

  

• Figure 7: Comparison of Evaluation metrics for different models and different datasets

  

• Figure 8: Actual vs. Predicted data for different models for AEP dataset.

  (a) Proposed REDf model, (b) SVR Model, (c) Facebook Prophet model, and (d) RFR model

  

• Figure 9: Loss curves of training and validation phases for the proposed REDf model in different datasets.

  (a) represents the curves for AEP dataset, (b) represents COMED dataset, (c) represents DAYTON dataset, and (d) represents PJME dataset

  

• Figure 10: Cross-validation results of the REDf model for the AEP dataset. Each row corresponds to a different cross-validation fold. The left column shows the training vs. validation loss curves. The middle column presents the actual vs. predicted energy demand for the first 100 test samples. The right column displays the error distributions

  

• Figure S1: Actual vs. Predicted data for different models for COMED dataset.

  (a) Proposed REDf model, (b) SVR Model, (c) Facebook Prophet model, and (d) RFR model

  

• Figure S2: Actual vs. Predicted data for different models for DAYTON dataset.

  (a) Proposed REDf model, (b) SVR Model, (c) Facebook Prophet model, and (d) RFR model

  

• Figure S3: Actual vs. Predicted data for different models for PJME dataset.

  (a) Proposed REDf model, (b) SVR Model, (c) Facebook Prophet model, and (d) RFR model

  

• Figure S4: Cross-validation results of the REDf model for the COMED dataset. Each row corresponds to a different cross-validation fold. The left column shows the training vs. validation loss curves. The middle column presents the actual vs. predicted energy demand for the first 100 test samples. The right column displays the error distributions

  

• Figure S5: Cross-validation results of the REDf model for the DAYTON dataset. Each row corresponds to a different cross-validation fold. The left column shows the training vs. validation loss curves. The middle column presents the actual vs. predicted energy demand for the first 100 test samples. The right column displays the error distributions

  

• Figure S6: Cross-validation results of the REDf model for the PJME dataset. Each row corresponds to a different cross-validation fold. The left column shows the training vs. validation loss curves. The middle column presents the actual vs. predicted energy demand for the first 100 test samples. The right column displays the error distributions