ABSTRACT
Over the past 15 years, there has been a significant change in the patterns towards deregulation and restructuring of the power industry with reduced electricity price. Therefore, this would encourage to forecast the electricity price in the deregulated power market with accurate model such that the generating companies are benefitted. The aim of this topic is to predict the electricity price, 24 hours in advance, on a day-head basis. Effectively, forecasting energy prices has useful applications, such as maximizing energy storage and allowing building flexibility on the demand side. Nowadays, deep learning methods like CNN (convolutional neural network) are used having image data. However, the RNN (recurrent neural network) and the type of RNN, namely LSTM (long short-term memory), are chosen as price forecaster. In this work, hourly electricity prices are collected from Nordic pool 2013–2019 data for price forecasting. For the case study, exogenous variables such as day of the week, hours of the day, temperature, oil prices, natural gas prices, coal prices, and historical price are considered. The upcoming 24-hour prices is predicted recursively. For validation of the pricing results, mean absolute percentage error (MAPE) and root mean square error (RMSE) are considered for the forecasting accuracy.
