Factor-augmented functional regression with an application to electricity price curve forecasting
-
Series
-
Speaker(s)Sven Otto (University of Cologne, Germany)
-
FieldEconometrics
-
LocationErasmus University Rotterdam, Mandeville building, room T3-24
Rotterdam -
Date and time
December 14, 2023
12:00 - 13:00
Abstract
We propose a function-on-function regression model for time-dependent curve data that is consistently estimated by imposing factor structures on the regressors. A novel integral operator $D$ identifies the predictive low-dimensional component with associated factors that are guaranteed to be correlated with the dependent variable. In order to consistently estimate the correct number of factors for each regressor, we introduce a functional eigenvalue difference test. The model is applied to forecast electricity price curves on three different energy markets. We show that the prediction accuracy of the factor-augmented functional regression is comparable to popular machine learning approaches while it provides interpretable insights into correlation structures of electricity prices.