Short Term Electricity Price Forecasting by Optimized LSTM Model of Deep Learning with Genetic Algorithm

The complexity of competitive markets, which are characterised by high volatility and frequent price spikes, makes forecasting power prices a challenging task. This study used various statistical methods to forecast day-ahead prices, however, statistical methods don’t provide good results for non-linear and high-volatility data. Thus to deal with the volatility in electricity prices, advanced deep-learning techniques have been used where LSTM model provides significantly better results. Furthermore, hyperparameter optimization is done using Genetic algorithm optimization which optimizes the hyperparameters of the LSTM model and enhances its performance. Forecasting performance is here evaluated by three well-known error metrics: Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Root Mean Square Error (RMSE).