Authors: José D. Bermúdez
Addresses: Department of Statistics and Operational Research, University of Valencia, Dr. Moliner, 50, E-46100 Burjassot, Spain
Abstract: Exponential smoothing methods are widely used as forecasting techniques in industry and business. Their usual formulation, however, does not allow covariates to be used for introducing extra information into the forecasting process. In this paper, we analyse an extension of the exponential smoothing formulation that allows the use of covariates and the joint estimation of all the unknowns in the model, which improves the forecasting results. The whole procedure is detailed with a real example on forecasting the daily demand for electricity in Spain. The time series of daily electricity demand contains two seasonal patterns: here the within-week seasonal cycle is modelled as usual in exponential smoothing, while the within-year cycle is modelled using covariates, specifically two harmonic explanatory variables. Calendar effects, such as national and local holidays and vacation periods, are also introduced using covariates. [Received 28 September 2010; Revised 6 March 2011, 2 October 2011; Accepted 16 October 2011]
Keywords: covariates; dynamic linear modelling; electricity demand; electricity forecasting; energy forecasting; exponential smoothing; forecasting practice; time series; Spain; calendar effects.
European Journal of Industrial Engineering, 2013 Vol.7 No.3, pp.333 - 349
Received: 08 May 2021
Accepted: 12 May 2021
Published online: 22 May 2013 *