Intelligent Energy Price Forecasting using Deep Learning

Energy Price forecasting is important towards meeting the demand of consumers and accordingly bring the consumers and utility play part in efficient usage of energy and generation resulting in reduced pricing. Previous works proposed machine learning technique on large data set with the predicted parameters such as price, energy, and demand for accurate predictions. However, forecasting on a country wide dataset with several regions remains challenging due to the complex dataset. In this study, two methods have been applied namely ARIMA and LSTM in an ensemble fashion on the AEMO Average Price dataset which consists of five regions over a period of more than two decades to predict the average RRP (Average spot price. The results obtained showed that the proposed LSTM method outperforms the ARIMA model.