Fast computation of reconciled forecasts for hierarchical and grouped time series

Robin John Hyndman, Alan J Lee, Yiru Earo Wang

Research output: Contribution to journalArticleResearchpeer-review

Abstract

It is shown that the least squares approach to reconciling hierarchical time series forecasts can be extended to much more general collections of time series with aggregation constraints. The constraints arise due to the need for forecasts of collections of time series to add up in the same way as the observed time series. It is also shown that the computations involved can be handled efficiently by exploiting the structure of the associated design matrix, or by using sparse matrix routines. The proposed algorithms make forecast reconciliation feasible in business applications involving very large numbers of time series.
Original languageEnglish
Pages (from-to)16 - 32
Number of pages17
JournalComputational Statistics and Data Analysis
Volume97
DOIs
Publication statusPublished - 2016

Keywords

  • combining forecasts
  • grouped time series
  • hierarchical time series
  • reconciling forecasts
  • weighted least squares

Cite this

@article{7c8411a8ddd549c780c59fea9ea17dbf,
title = "Fast computation of reconciled forecasts for hierarchical and grouped time series",
abstract = "It is shown that the least squares approach to reconciling hierarchical time series forecasts can be extended to much more general collections of time series with aggregation constraints. The constraints arise due to the need for forecasts of collections of time series to add up in the same way as the observed time series. It is also shown that the computations involved can be handled efficiently by exploiting the structure of the associated design matrix, or by using sparse matrix routines. The proposed algorithms make forecast reconciliation feasible in business applications involving very large numbers of time series.",
keywords = "combining forecasts, grouped time series, hierarchical time series, reconciling forecasts, weighted least squares",
author = "Hyndman, {Robin John} and Lee, {Alan J} and Wang, {Yiru Earo}",
year = "2016",
doi = "10.1016/j.csda.2015.11.007",
language = "English",
volume = "97",
pages = "16 -- 32",
journal = "Computational Statistics and Data Analysis",
issn = "0167-9473",
publisher = "Elsevier",

}

Fast computation of reconciled forecasts for hierarchical and grouped time series. / Hyndman, Robin John; Lee, Alan J; Wang, Yiru Earo.

In: Computational Statistics and Data Analysis, Vol. 97, 2016, p. 16 - 32.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Fast computation of reconciled forecasts for hierarchical and grouped time series

AU - Hyndman, Robin John

AU - Lee, Alan J

AU - Wang, Yiru Earo

PY - 2016

Y1 - 2016

N2 - It is shown that the least squares approach to reconciling hierarchical time series forecasts can be extended to much more general collections of time series with aggregation constraints. The constraints arise due to the need for forecasts of collections of time series to add up in the same way as the observed time series. It is also shown that the computations involved can be handled efficiently by exploiting the structure of the associated design matrix, or by using sparse matrix routines. The proposed algorithms make forecast reconciliation feasible in business applications involving very large numbers of time series.

AB - It is shown that the least squares approach to reconciling hierarchical time series forecasts can be extended to much more general collections of time series with aggregation constraints. The constraints arise due to the need for forecasts of collections of time series to add up in the same way as the observed time series. It is also shown that the computations involved can be handled efficiently by exploiting the structure of the associated design matrix, or by using sparse matrix routines. The proposed algorithms make forecast reconciliation feasible in business applications involving very large numbers of time series.

KW - combining forecasts

KW - grouped time series

KW - hierarchical time series

KW - reconciling forecasts

KW - weighted least squares

U2 - 10.1016/j.csda.2015.11.007

DO - 10.1016/j.csda.2015.11.007

M3 - Article

VL - 97

SP - 16

EP - 32

JO - Computational Statistics and Data Analysis

JF - Computational Statistics and Data Analysis

SN - 0167-9473

ER -