Computing electricity spot price prediction intervals using quantile regression and forecast averaging

Abstract

We examine possible accuracy gains from forecast averaging in the context of interval forecasts of electricity spot prices. First, we test whether constructing empirical prediction intervals (PI) from combined electricity spot price forecasts leads to better forecasts than those obtained from individual methods. Next, we propose a new method for constructing PI-Quantile Regression Averaging (QRA)-which utilizes the concept of quantile regression and a pool of point forecasts of individual (i.e. not combined) models. While the empirical PI from combined forecasts do not provide significant gains, the QRA-based PI are found to be more accurate than those of the best individual model-the smoothed nonparametric autoregressive model.

Editorial to the special issue on Applicable semiparametrics of computational statistics

“…Nowotarski and Weron (2015) investigate interval forecast techniques that are based on combined individual electricity spot price forecasts and quantile regression averaging (QRA). The authors suggest that in particular QRA-based forecasting yield promising results that provide accuracy gains over predictions from individual models.…”

Probabilistic Price Forecasting for Day-Ahead and Intraday Markets: Beyond the Statistical Model

“…The input data were zonal and system loads, lagged prices, standard deviation of the price during the previous day, maximum price variation from the previous day and calendar variables. Nowotarski and Weron proposed the method Quantile Regression Averaging (QRA) that applies linear quantile regression to a pool of individual point forecasts and showed that it outperformed the best individual model [21]. QRA was extended by Maciejowsk et al by using principal component analysis for selecting a subset of individual forecasts [22].…”

Predictive Densities for Day-Ahead Electricity Prices Using Time-Adaptive Quantile Regression

“…Besides, [19,20] present interesting methods for modeling the first four moments of conditional price densities with alternative probability density functions. More recently, novel nonparametric approaches were proposed based on quantile regression averaging, therefore allowing one to issue probabilistic forecasts in a forecast combination framework, e.g., [21,22].…”