Coherent probabilistic forecasts for hierarchical time series

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

Abstract

Many applications require forecasts for a hierarchy comprising a set of time scries along with aggregates of subsets of these scries. Hierarchical forecasting require not only good prediction accuracy at each level of the hierarchy, but also the coherency between different levels-the property that forecasts add up appropriately across the hierarchy. A fundamental limitation of prior research is the focus on forecasting the mean of each time series. We consider the situation where probabilistic forecasts are needed for each series in the hierarchy, and propose an algorithm to compute predictive distributions rather than mean forecasts only. Our algorithm has the advantage of synthesizing information from different levels in the hierarchy through a sparse forecast combination and a probabilistic hierarchical aggregation. We evaluate the accuracy of our forecasting algorithm on both simulated data and large-scale electricity smart meter data. The results show consistent performance gains compared to state-of-the art methods.

Original languageEnglish
Title of host publicationProceedings of the 34th International Conference on Machine Learning
EditorsDoina Precup, Yee Whye Teh
Place of PublicationMassachusetts USA
PublisherProceedings of Machine Learning Research (PMLR)
Pages3348-3357
Number of pages10
Publication statusPublished - 1 Jan 2017
EventInternational Conference on Machine Learning 2017 - International Convention Centre , Sydney , Australia
Duration: 6 Aug 201711 Aug 2018
Conference number: 34th
https://icml.cc/Conferences/2017
https://2017.icml.cc/

Publication series

NameProceedings of Machine Learning Research
Volume70
ISSN (Print)1938-7228

Conference

ConferenceInternational Conference on Machine Learning 2017
Abbreviated titleICML 2017
CountryAustralia
CitySydney
Period6/08/1711/08/18
Internet address

Cite this

Taieb, S. B., Taylor, J. W., & Hyndman, R. J. (2017). Coherent probabilistic forecasts for hierarchical time series. In D. Precup, & Y. W. Teh (Eds.), Proceedings of the 34th International Conference on Machine Learning (pp. 3348-3357). (Proceedings of Machine Learning Research; Vol. 70). Massachusetts USA: Proceedings of Machine Learning Research (PMLR).
Taieb, Souhaib Ben ; Taylor, James W. ; Hyndman, Rob J. / Coherent probabilistic forecasts for hierarchical time series. Proceedings of the 34th International Conference on Machine Learning. editor / Doina Precup ; Yee Whye Teh. Massachusetts USA : Proceedings of Machine Learning Research (PMLR), 2017. pp. 3348-3357 (Proceedings of Machine Learning Research).
@inproceedings{fd0bb56b4703431c9dcd6b7b4b46cb5a,
title = "Coherent probabilistic forecasts for hierarchical time series",
abstract = "Many applications require forecasts for a hierarchy comprising a set of time scries along with aggregates of subsets of these scries. Hierarchical forecasting require not only good prediction accuracy at each level of the hierarchy, but also the coherency between different levels-the property that forecasts add up appropriately across the hierarchy. A fundamental limitation of prior research is the focus on forecasting the mean of each time series. We consider the situation where probabilistic forecasts are needed for each series in the hierarchy, and propose an algorithm to compute predictive distributions rather than mean forecasts only. Our algorithm has the advantage of synthesizing information from different levels in the hierarchy through a sparse forecast combination and a probabilistic hierarchical aggregation. We evaluate the accuracy of our forecasting algorithm on both simulated data and large-scale electricity smart meter data. The results show consistent performance gains compared to state-of-the art methods.",
author = "Taieb, {Souhaib Ben} and Taylor, {James W.} and Hyndman, {Rob J}",
year = "2017",
month = "1",
day = "1",
language = "English",
series = "Proceedings of Machine Learning Research",
publisher = "Proceedings of Machine Learning Research (PMLR)",
pages = "3348--3357",
editor = "Doina Precup and Teh, {Yee Whye}",
booktitle = "Proceedings of the 34th International Conference on Machine Learning",

}

Taieb, SB, Taylor, JW & Hyndman, RJ 2017, Coherent probabilistic forecasts for hierarchical time series. in D Precup & YW Teh (eds), Proceedings of the 34th International Conference on Machine Learning. Proceedings of Machine Learning Research, vol. 70, Proceedings of Machine Learning Research (PMLR), Massachusetts USA, pp. 3348-3357, International Conference on Machine Learning 2017, Sydney , Australia, 6/08/17.

Coherent probabilistic forecasts for hierarchical time series. / Taieb, Souhaib Ben; Taylor, James W.; Hyndman, Rob J.

Proceedings of the 34th International Conference on Machine Learning. ed. / Doina Precup; Yee Whye Teh. Massachusetts USA : Proceedings of Machine Learning Research (PMLR), 2017. p. 3348-3357 (Proceedings of Machine Learning Research; Vol. 70).

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

TY - GEN

T1 - Coherent probabilistic forecasts for hierarchical time series

AU - Taieb, Souhaib Ben

AU - Taylor, James W.

AU - Hyndman, Rob J

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Many applications require forecasts for a hierarchy comprising a set of time scries along with aggregates of subsets of these scries. Hierarchical forecasting require not only good prediction accuracy at each level of the hierarchy, but also the coherency between different levels-the property that forecasts add up appropriately across the hierarchy. A fundamental limitation of prior research is the focus on forecasting the mean of each time series. We consider the situation where probabilistic forecasts are needed for each series in the hierarchy, and propose an algorithm to compute predictive distributions rather than mean forecasts only. Our algorithm has the advantage of synthesizing information from different levels in the hierarchy through a sparse forecast combination and a probabilistic hierarchical aggregation. We evaluate the accuracy of our forecasting algorithm on both simulated data and large-scale electricity smart meter data. The results show consistent performance gains compared to state-of-the art methods.

AB - Many applications require forecasts for a hierarchy comprising a set of time scries along with aggregates of subsets of these scries. Hierarchical forecasting require not only good prediction accuracy at each level of the hierarchy, but also the coherency between different levels-the property that forecasts add up appropriately across the hierarchy. A fundamental limitation of prior research is the focus on forecasting the mean of each time series. We consider the situation where probabilistic forecasts are needed for each series in the hierarchy, and propose an algorithm to compute predictive distributions rather than mean forecasts only. Our algorithm has the advantage of synthesizing information from different levels in the hierarchy through a sparse forecast combination and a probabilistic hierarchical aggregation. We evaluate the accuracy of our forecasting algorithm on both simulated data and large-scale electricity smart meter data. The results show consistent performance gains compared to state-of-the art methods.

UR - http://www.scopus.com/inward/record.url?scp=85048460620&partnerID=8YFLogxK

M3 - Conference Paper

T3 - Proceedings of Machine Learning Research

SP - 3348

EP - 3357

BT - Proceedings of the 34th International Conference on Machine Learning

A2 - Precup, Doina

A2 - Teh, Yee Whye

PB - Proceedings of Machine Learning Research (PMLR)

CY - Massachusetts USA

ER -

Taieb SB, Taylor JW, Hyndman RJ. Coherent probabilistic forecasts for hierarchical time series. In Precup D, Teh YW, editors, Proceedings of the 34th International Conference on Machine Learning. Massachusetts USA: Proceedings of Machine Learning Research (PMLR). 2017. p. 3348-3357. (Proceedings of Machine Learning Research).