Probabilistic short-term load forecasting with conditional mean-variance and quantile regression models

C. Bikcora, L. Verheijen, S. Weiland

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

Abstract

For the day-ahead density forecasting of electricity load, this paper proposes the combination of the autoregressive moving average (ARMA) model and the generalized autoregressive conditional heteroskedasticity (GARCH) model, with both of them admitting exogenous inputs. This composite structure on the conditional mean and variance is referred to as the ARMAX-GARCHX model. As an alternative to its estimation by means of log-likelihood maximization, approaches based on iterative least-squares (ILS) and nonlinear least-squares (NLS) are considered. Apart from the ARMAX-GARCHX model, quantile regression models (QRMs) are also tested in forecasting where a wide range of quantiles are separately modeled to approximate a density. Phase currents of several low voltage transformer cables from the Netherlands are forecasted to compare the performances, and as the probabilistic evaluation criterion, the continuous ranked probability score is used. As an outline of the results, the ARMAX-GARCHX model outperformed QRMs and among its estimation techniques, the likelihood-based approach had the best performance, though the differences in the errors are often minor. Thus, owing to its computational simplicity, the ILS solution can be a valuable option when processing large batches of data in practice.
LanguageEnglish
Title of host publicationProceedings of European Control Conference (ECC), Aalborg, Denmark, Jun. 2016
PublisherInstitute of Electrical and Electronics Engineers
Pages1123-1128
Number of pages6
ISBN (Electronic)978-1-5090-2591-6
DOIs
StatePublished - Jun 2016
Event15th European Control Conference (ECC 2016) - Aalborg, Denmark
Duration: 29 Jun 20161 Jul 2016
Conference number: 15
http://www.ecc16.eu/index.shtml
http://www.ecc16.eu/index.shtml

Conference

Conference15th European Control Conference (ECC 2016)
Abbreviated titleECC 2016
CountryDenmark
CityAalborg
Period29/06/161/07/16
Internet address

Fingerprint

Composite structures
Cables
Electricity
Electric potential
Processing

Cite this

Bikcora, C., Verheijen, L., & Weiland, S. (2016). Probabilistic short-term load forecasting with conditional mean-variance and quantile regression models. In Proceedings of European Control Conference (ECC), Aalborg, Denmark, Jun. 2016 (pp. 1123-1128). [Paper ThA8.1] Institute of Electrical and Electronics Engineers. DOI: 10.1109/ECC.2016.7810440
Bikcora, C. ; Verheijen, L. ; Weiland, S./ Probabilistic short-term load forecasting with conditional mean-variance and quantile regression models. Proceedings of European Control Conference (ECC), Aalborg, Denmark, Jun. 2016. Institute of Electrical and Electronics Engineers, 2016. pp. 1123-1128
@inproceedings{d44cd5cf599b4badae54976a1f70d1de,
title = "Probabilistic short-term load forecasting with conditional mean-variance and quantile regression models",
abstract = "For the day-ahead density forecasting of electricity load, this paper proposes the combination of the autoregressive moving average (ARMA) model and the generalized autoregressive conditional heteroskedasticity (GARCH) model, with both of them admitting exogenous inputs. This composite structure on the conditional mean and variance is referred to as the ARMAX-GARCHX model. As an alternative to its estimation by means of log-likelihood maximization, approaches based on iterative least-squares (ILS) and nonlinear least-squares (NLS) are considered. Apart from the ARMAX-GARCHX model, quantile regression models (QRMs) are also tested in forecasting where a wide range of quantiles are separately modeled to approximate a density. Phase currents of several low voltage transformer cables from the Netherlands are forecasted to compare the performances, and as the probabilistic evaluation criterion, the continuous ranked probability score is used. As an outline of the results, the ARMAX-GARCHX model outperformed QRMs and among its estimation techniques, the likelihood-based approach had the best performance, though the differences in the errors are often minor. Thus, owing to its computational simplicity, the ILS solution can be a valuable option when processing large batches of data in practice.",
author = "C. Bikcora and L. Verheijen and S. Weiland",
year = "2016",
month = "6",
doi = "10.1109/ECC.2016.7810440",
language = "English",
pages = "1123--1128",
booktitle = "Proceedings of European Control Conference (ECC), Aalborg, Denmark, Jun. 2016",
publisher = "Institute of Electrical and Electronics Engineers",
address = "United States",

}

Bikcora, C, Verheijen, L & Weiland, S 2016, Probabilistic short-term load forecasting with conditional mean-variance and quantile regression models. in Proceedings of European Control Conference (ECC), Aalborg, Denmark, Jun. 2016., Paper ThA8.1, Institute of Electrical and Electronics Engineers, pp. 1123-1128, 15th European Control Conference (ECC 2016), Aalborg, Denmark, 29/06/16. DOI: 10.1109/ECC.2016.7810440

Probabilistic short-term load forecasting with conditional mean-variance and quantile regression models. / Bikcora, C.; Verheijen, L.; Weiland, S.

Proceedings of European Control Conference (ECC), Aalborg, Denmark, Jun. 2016. Institute of Electrical and Electronics Engineers, 2016. p. 1123-1128 Paper ThA8.1.

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Probabilistic short-term load forecasting with conditional mean-variance and quantile regression models

AU - Bikcora,C.

AU - Verheijen,L.

AU - Weiland,S.

PY - 2016/6

Y1 - 2016/6

N2 - For the day-ahead density forecasting of electricity load, this paper proposes the combination of the autoregressive moving average (ARMA) model and the generalized autoregressive conditional heteroskedasticity (GARCH) model, with both of them admitting exogenous inputs. This composite structure on the conditional mean and variance is referred to as the ARMAX-GARCHX model. As an alternative to its estimation by means of log-likelihood maximization, approaches based on iterative least-squares (ILS) and nonlinear least-squares (NLS) are considered. Apart from the ARMAX-GARCHX model, quantile regression models (QRMs) are also tested in forecasting where a wide range of quantiles are separately modeled to approximate a density. Phase currents of several low voltage transformer cables from the Netherlands are forecasted to compare the performances, and as the probabilistic evaluation criterion, the continuous ranked probability score is used. As an outline of the results, the ARMAX-GARCHX model outperformed QRMs and among its estimation techniques, the likelihood-based approach had the best performance, though the differences in the errors are often minor. Thus, owing to its computational simplicity, the ILS solution can be a valuable option when processing large batches of data in practice.

AB - For the day-ahead density forecasting of electricity load, this paper proposes the combination of the autoregressive moving average (ARMA) model and the generalized autoregressive conditional heteroskedasticity (GARCH) model, with both of them admitting exogenous inputs. This composite structure on the conditional mean and variance is referred to as the ARMAX-GARCHX model. As an alternative to its estimation by means of log-likelihood maximization, approaches based on iterative least-squares (ILS) and nonlinear least-squares (NLS) are considered. Apart from the ARMAX-GARCHX model, quantile regression models (QRMs) are also tested in forecasting where a wide range of quantiles are separately modeled to approximate a density. Phase currents of several low voltage transformer cables from the Netherlands are forecasted to compare the performances, and as the probabilistic evaluation criterion, the continuous ranked probability score is used. As an outline of the results, the ARMAX-GARCHX model outperformed QRMs and among its estimation techniques, the likelihood-based approach had the best performance, though the differences in the errors are often minor. Thus, owing to its computational simplicity, the ILS solution can be a valuable option when processing large batches of data in practice.

U2 - 10.1109/ECC.2016.7810440

DO - 10.1109/ECC.2016.7810440

M3 - Conference contribution

SP - 1123

EP - 1128

BT - Proceedings of European Control Conference (ECC), Aalborg, Denmark, Jun. 2016

PB - Institute of Electrical and Electronics Engineers

ER -

Bikcora C, Verheijen L, Weiland S. Probabilistic short-term load forecasting with conditional mean-variance and quantile regression models. In Proceedings of European Control Conference (ECC), Aalborg, Denmark, Jun. 2016. Institute of Electrical and Electronics Engineers. 2016. p. 1123-1128. Paper ThA8.1. Available from, DOI: 10.1109/ECC.2016.7810440