Electric Power Load Forecasting Based on Multivariate LSTM Neural Network Using Bayesian Optimization
Mohammad Munem, T. M. Rubaith Bashar, Mehedi Hasan Roni, Munem Shahriar, Tasnim Binte Shawkat, Habibur Rahaman
Abstract
With rapid growth and development around the world, electricity consumption is increasing day by day. As the production and consumption of electricity is simultaneous, an electric power load forecasting technique with higher accuracy can play a pivotal role in a stable and effective power supply system. In this paper, a multivariate Bayesian optimization based Long short-term memory (LSTM) neural network is proposed to forecast the residential electric power load for the upcoming hour. Bayesian optimization algorithm is conducted to select the best-fitted hyperparameter values since deep learning networks are associated with different hyperparameters which play a vital role in the performance of a network architecture. Our proposed Bayesian optimized LSTM neural network has obtained almost perfect prediction performance and it surpasses the other established model such as convolutional neural network (CNN), artificial neural network (ANN) and support vector machine (SVM) where mean absolute error (MAE), root mean squared error (RMSE) and mean squared error (MSE) are found 0.39, 0.54 and 0.29 respectively for the individual household power consumption dataset.