This article proposes a semiparametric single-index model for short-term forecasting day-ahead electricity prices. The approach captures the dependency of electricity prices on covariates, such as demand for electricity, amount of energy produced by intermittent sources, and weather-dependent variables. To obtain parsimonious models, principal component analysis is used for dimension reduction. The approach is tested on two data sets from different markets and its performance is analyzed in terms of fit, forecast quality, and computational efficiency. The results are encouraging, in that the proposed method leads to a good in-sample fit and performs well out-of-sample compared with four benchmark models, including a SARIMA model as well as a functional nonparametric regression approach recently proposed in the literature.     «

This article proposes a semiparametric single-index model for short-term forecasting day-ahead electricity prices. The approach captures the dependency of electricity prices on covariates, such as demand for electricity, amount of energy produced by intermittent sources, and weather-dependent variables. To obtain parsimonious models, principal component analysis is used for dimension reduction. The approach is tested on two data sets from different markets and its performance is analyzed in term...     »