Sparse and smooth adjustments for coherent forecasts in temporal aggregation of time series

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

Abstract

Independent forecasts obtained from different temporal aggregates of a given time series may not be mutually consistent. State-of-the art forecasting methods usually apply adjustments on the individual level forecasts to satisfy the aggregation constraints. These adjustments require the estimation of the covariance between the individual forecast errors at all aggregation levels. In order to keep a maximum number of individual forecasts unaffected by estimation errors, we propose a new forecasting algorithm that provides sparse
and smooth adjustments while still preserving the aggregation constraints. The algorithm computes the revised forecasts by solving a generalized lasso problem. It is shown that it not only provides accurate forecasts, but also applies a significantly smaller number of adjustments to the base forecasts in a large-scale smart meter dataset.
Original languageEnglish
Title of host publicationNIPS 2016 Time Series Workshop, 09 December 2016, Barcelona, Spain
EditorsOren Anava, Azadeh Khaleghi, Marco Cuturi, Vitaly Kuznetsov, Alexander Rakhlin
Place of PublicationUSA
PublisherProceedings of Machine Learning Research (PMLR)
Number of pages11
Volume55
Publication statusPublished - 2016
EventNIPS Time Series Workshop 2016 - Centre Convencions Internacional Barcelona, Barcelona, Spain
Duration: 9 Dec 20169 Dec 2016
https://sites.google.com/site/nipsts2016/home

Workshop

WorkshopNIPS Time Series Workshop 2016
CountrySpain
CityBarcelona
Period9/12/169/12/16
Internet address

Cite this

Ben Taieb, S. (2016). Sparse and smooth adjustments for coherent forecasts in temporal aggregation of time series. In O. Anava, A. Khaleghi, M. Cuturi, V. Kuznetsov, & A. Rakhlin (Eds.), NIPS 2016 Time Series Workshop, 09 December 2016, Barcelona, Spain (Vol. 55). USA: Proceedings of Machine Learning Research (PMLR).
Ben Taieb, Souhaib. / Sparse and smooth adjustments for coherent forecasts in temporal aggregation of time series. NIPS 2016 Time Series Workshop, 09 December 2016, Barcelona, Spain. editor / Oren Anava ; Azadeh Khaleghi ; Marco Cuturi ; Vitaly Kuznetsov ; Alexander Rakhlin. Vol. 55 USA : Proceedings of Machine Learning Research (PMLR), 2016.
@inproceedings{0d012bdd201340bc8697a92c8220cb93,
title = "Sparse and smooth adjustments for coherent forecasts in temporal aggregation of time series",
abstract = "Independent forecasts obtained from different temporal aggregates of a given time series may not be mutually consistent. State-of-the art forecasting methods usually apply adjustments on the individual level forecasts to satisfy the aggregation constraints. These adjustments require the estimation of the covariance between the individual forecast errors at all aggregation levels. In order to keep a maximum number of individual forecasts unaffected by estimation errors, we propose a new forecasting algorithm that provides sparseand smooth adjustments while still preserving the aggregation constraints. The algorithm computes the revised forecasts by solving a generalized lasso problem. It is shown that it not only provides accurate forecasts, but also applies a significantly smaller number of adjustments to the base forecasts in a large-scale smart meter dataset.",
author = "{Ben Taieb}, Souhaib",
year = "2016",
language = "English",
volume = "55",
editor = "Oren Anava and Azadeh Khaleghi and Marco Cuturi and Vitaly Kuznetsov and Alexander Rakhlin",
booktitle = "NIPS 2016 Time Series Workshop, 09 December 2016, Barcelona, Spain",
publisher = "Proceedings of Machine Learning Research (PMLR)",

}

Ben Taieb, S 2016, Sparse and smooth adjustments for coherent forecasts in temporal aggregation of time series. in O Anava, A Khaleghi, M Cuturi, V Kuznetsov & A Rakhlin (eds), NIPS 2016 Time Series Workshop, 09 December 2016, Barcelona, Spain. vol. 55, Proceedings of Machine Learning Research (PMLR), USA, NIPS Time Series Workshop 2016, Barcelona, Spain, 9/12/16.

Sparse and smooth adjustments for coherent forecasts in temporal aggregation of time series. / Ben Taieb, Souhaib.

NIPS 2016 Time Series Workshop, 09 December 2016, Barcelona, Spain. ed. / Oren Anava; Azadeh Khaleghi; Marco Cuturi; Vitaly Kuznetsov; Alexander Rakhlin. Vol. 55 USA : Proceedings of Machine Learning Research (PMLR), 2016.

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

TY - GEN

T1 - Sparse and smooth adjustments for coherent forecasts in temporal aggregation of time series

AU - Ben Taieb, Souhaib

PY - 2016

Y1 - 2016

N2 - Independent forecasts obtained from different temporal aggregates of a given time series may not be mutually consistent. State-of-the art forecasting methods usually apply adjustments on the individual level forecasts to satisfy the aggregation constraints. These adjustments require the estimation of the covariance between the individual forecast errors at all aggregation levels. In order to keep a maximum number of individual forecasts unaffected by estimation errors, we propose a new forecasting algorithm that provides sparseand smooth adjustments while still preserving the aggregation constraints. The algorithm computes the revised forecasts by solving a generalized lasso problem. It is shown that it not only provides accurate forecasts, but also applies a significantly smaller number of adjustments to the base forecasts in a large-scale smart meter dataset.

AB - Independent forecasts obtained from different temporal aggregates of a given time series may not be mutually consistent. State-of-the art forecasting methods usually apply adjustments on the individual level forecasts to satisfy the aggregation constraints. These adjustments require the estimation of the covariance between the individual forecast errors at all aggregation levels. In order to keep a maximum number of individual forecasts unaffected by estimation errors, we propose a new forecasting algorithm that provides sparseand smooth adjustments while still preserving the aggregation constraints. The algorithm computes the revised forecasts by solving a generalized lasso problem. It is shown that it not only provides accurate forecasts, but also applies a significantly smaller number of adjustments to the base forecasts in a large-scale smart meter dataset.

M3 - Conference Paper

VL - 55

BT - NIPS 2016 Time Series Workshop, 09 December 2016, Barcelona, Spain

A2 - Anava, Oren

A2 - Khaleghi, Azadeh

A2 - Cuturi, Marco

A2 - Kuznetsov, Vitaly

A2 - Rakhlin, Alexander

PB - Proceedings of Machine Learning Research (PMLR)

CY - USA

ER -

Ben Taieb S. Sparse and smooth adjustments for coherent forecasts in temporal aggregation of time series. In Anava O, Khaleghi A, Cuturi M, Kuznetsov V, Rakhlin A, editors, NIPS 2016 Time Series Workshop, 09 December 2016, Barcelona, Spain. Vol. 55. USA: Proceedings of Machine Learning Research (PMLR). 2016