A Comparative Study on Deep Learning Methods for Forecasting Load in Smart Grid

This study compares deep learning models such as RNN-LSTM, DenseNet, and ResNet, along with a hybrid model combining SVM, XGBoost, and DTC, for load forecasting in power systems. The results demonstrate the strengths and weaknesses of each approach, with RNN-LSTM capturing temporal patterns effectively, DenseNet and ResNet excelling in feature extraction, and the hybrid model offering a balanced combination of precision and effectiveness. These findings offer valuable insights for power system operators and researchers in selecting suitable load forecasting models.