Forecasting Electricity Prices from European Single Day-ahead Coupling using Artificial Neural Networks

The electricity wholesale market is one of the most complex economic institutions. Prices on this fast-moving market depend on volatile factors such as weather conditions or abrupt changes in demand. Furthermore, the level of physical integration – and thus interdependence - between countries in Europe is increasing. As such, the electricity wholesale market demands most accurate price forecasting models to provide vital competitive advantages to actors on the market. This paper aims to develop and to evaluate the performance of different deep neural network models for the purpose of electricity price forecasting, focusing on the Austrian market. To incorporate the effect of market coupling in Europe, data from various other European countries are considered in the models.