 Our proposed approach combines machine learning and LSTM-based neural networks to construct forecasting models for short to medium-term aggregate load forecasting. We first train multiple linear and non-linear machine learning algorithms on the dataset and select the best one as a baseline. Then we use wrapper and embedded feature selection methods to choose the most relevant features from the dataset. Finally, we optimize the parameters of the LSTM model using genetic algorithm, GA. This combination of techniques resulted in a significant improvement in forecasting accuracy compared to other approaches. This article was authored by Sula Baoktiff, Ali Faiz, Ali Auni, and others.