Electricity Load Forecasting: A New Hybrid Model Combining Wavelet Transform, Genetic Algorithm and LSTM Models

Abstract

Electricity load forecasting plays an important role in energy planning. In this paper artificial intelligence and metaheuristic algorithms are combined to develop an electricity load forecasting model. In this regard, the Long Short-Term Memory (LSTM) neural network is used as well as genetic algorithm (GA) for optimizing the parameters. Furthermore, in order to increase the forecasting accuracy, the input electricity load demand is decomposed on the basis of wavelet transform (WT). The comparison of the proposed and available models shows the superiority of the proposed model rather than the other ones. The evaluation criteria include Mean Absolute Percent Error (MAPE), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE) indicators.