Forecasting uncertainty in electricity smart meter data by boosting additive quantile regression

Souhaib Ben Taieb, Raphael Huser, Rob J. Hyndman, Marc G. Genton

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Smart electricity meters are currently deployed in millions of households to collect detailed individual electricity consumption data. Compared with traditional electricity data based on aggregated consumption, smart meter data are much more volatile and less predictable. There is a need within the energy industry for probabilistic forecasts of household electricity consumption to quantify the uncertainty of future electricity demand in order to undertake appropriate planning of generation and distribution. We propose to estimate an additive quantile regression model for a set of quantiles of the future distribution using a boosting procedure. By doing so, we can benefit from flexible and interpretable models, which include an automatic variable selection. We compare our approach with three benchmark methods on both aggregated and disaggregated scales using a smart meter data set collected from 3639 households in Ireland at 30-min intervals over a period of 1.5 years. The empirical results demonstrate that our approach based on quantile regression provides better forecast accuracy for disaggregated demand, while the traditional approach based on a normality assumption (possibly after an appropriate Box-Cox transformation) is a better approximation for aggregated demand. These results are particularly useful since more energy data will become available at the disaggregated level in the future.

Original languageEnglish
Article number7423794
Pages (from-to)2448-2455
Number of pages8
JournalIEEE Transactions on Smart Grid
Volume7
Issue number5
DOIs
Publication statusPublished - 1 Sep 2016

Keywords

  • gradient boosting
  • Probabilistic load forecasting
  • quantile regression
  • smart meters

Cite this

@article{b5959c5e02b245aabeebc40575738171,
title = "Forecasting uncertainty in electricity smart meter data by boosting additive quantile regression",
abstract = "Smart electricity meters are currently deployed in millions of households to collect detailed individual electricity consumption data. Compared with traditional electricity data based on aggregated consumption, smart meter data are much more volatile and less predictable. There is a need within the energy industry for probabilistic forecasts of household electricity consumption to quantify the uncertainty of future electricity demand in order to undertake appropriate planning of generation and distribution. We propose to estimate an additive quantile regression model for a set of quantiles of the future distribution using a boosting procedure. By doing so, we can benefit from flexible and interpretable models, which include an automatic variable selection. We compare our approach with three benchmark methods on both aggregated and disaggregated scales using a smart meter data set collected from 3639 households in Ireland at 30-min intervals over a period of 1.5 years. The empirical results demonstrate that our approach based on quantile regression provides better forecast accuracy for disaggregated demand, while the traditional approach based on a normality assumption (possibly after an appropriate Box-Cox transformation) is a better approximation for aggregated demand. These results are particularly useful since more energy data will become available at the disaggregated level in the future.",
keywords = "gradient boosting, Probabilistic load forecasting, quantile regression, smart meters",
author = "{Ben Taieb}, Souhaib and Raphael Huser and Hyndman, {Rob J.} and Genton, {Marc G.}",
year = "2016",
month = "9",
day = "1",
doi = "10.1109/TSG.2016.2527820",
language = "English",
volume = "7",
pages = "2448--2455",
journal = "IEEE Transactions on Smart Grid",
issn = "1949-3053",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
number = "5",

}

Forecasting uncertainty in electricity smart meter data by boosting additive quantile regression. / Ben Taieb, Souhaib; Huser, Raphael; Hyndman, Rob J.; Genton, Marc G.

In: IEEE Transactions on Smart Grid, Vol. 7, No. 5, 7423794, 01.09.2016, p. 2448-2455.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Forecasting uncertainty in electricity smart meter data by boosting additive quantile regression

AU - Ben Taieb, Souhaib

AU - Huser, Raphael

AU - Hyndman, Rob J.

AU - Genton, Marc G.

PY - 2016/9/1

Y1 - 2016/9/1

N2 - Smart electricity meters are currently deployed in millions of households to collect detailed individual electricity consumption data. Compared with traditional electricity data based on aggregated consumption, smart meter data are much more volatile and less predictable. There is a need within the energy industry for probabilistic forecasts of household electricity consumption to quantify the uncertainty of future electricity demand in order to undertake appropriate planning of generation and distribution. We propose to estimate an additive quantile regression model for a set of quantiles of the future distribution using a boosting procedure. By doing so, we can benefit from flexible and interpretable models, which include an automatic variable selection. We compare our approach with three benchmark methods on both aggregated and disaggregated scales using a smart meter data set collected from 3639 households in Ireland at 30-min intervals over a period of 1.5 years. The empirical results demonstrate that our approach based on quantile regression provides better forecast accuracy for disaggregated demand, while the traditional approach based on a normality assumption (possibly after an appropriate Box-Cox transformation) is a better approximation for aggregated demand. These results are particularly useful since more energy data will become available at the disaggregated level in the future.

AB - Smart electricity meters are currently deployed in millions of households to collect detailed individual electricity consumption data. Compared with traditional electricity data based on aggregated consumption, smart meter data are much more volatile and less predictable. There is a need within the energy industry for probabilistic forecasts of household electricity consumption to quantify the uncertainty of future electricity demand in order to undertake appropriate planning of generation and distribution. We propose to estimate an additive quantile regression model for a set of quantiles of the future distribution using a boosting procedure. By doing so, we can benefit from flexible and interpretable models, which include an automatic variable selection. We compare our approach with three benchmark methods on both aggregated and disaggregated scales using a smart meter data set collected from 3639 households in Ireland at 30-min intervals over a period of 1.5 years. The empirical results demonstrate that our approach based on quantile regression provides better forecast accuracy for disaggregated demand, while the traditional approach based on a normality assumption (possibly after an appropriate Box-Cox transformation) is a better approximation for aggregated demand. These results are particularly useful since more energy data will become available at the disaggregated level in the future.

KW - gradient boosting

KW - Probabilistic load forecasting

KW - quantile regression

KW - smart meters

UR - http://www.scopus.com/inward/record.url?scp=84960193121&partnerID=8YFLogxK

U2 - 10.1109/TSG.2016.2527820

DO - 10.1109/TSG.2016.2527820

M3 - Article

VL - 7

SP - 2448

EP - 2455

JO - IEEE Transactions on Smart Grid

JF - IEEE Transactions on Smart Grid

SN - 1949-3053

IS - 5

M1 - 7423794

ER -