Probabilistic forecasting of electricity spot prices using Factor Quantile Regression Averaging

We examine possible accuracy gains from using factor models, quantile regression and forecast averaging for computing interval forecasts of electricity spot prices. We extend the Quantile Regression Averaging (QRA) approach of Nowotarski and Weron (2014) and use principal component analysis to automate the selection process from among a large set of individual forecasting models available for averaging. We show that the resulting Factor Quantile Regression Averaging (FQRA) approach performs very well for price (and load) data from the British power market. In terms of unconditional coverage, conditional coverage and the Winkler score, we find the FQRA-implied prediction intervals to be more accurate than those of the benchmark ARX model and the QRA approach.