pith. sign in

arxiv: 2606.23009 · v1 · pith:VC6QLCSDnew · submitted 2026-06-22 · 📊 stat.ME

Hierarchical Bayes meets hierarchical forecasting: A flexible framework for level-focused forecasts

Arwen Nugteren , Mahdi Abolghasemi , Kerrie Mengersen , Christopher Drovandi This is my paper

Pith reviewed 2026-06-26 07:43 UTC · model grok-4.3

classification 📊 stat.ME
keywords hierarchical forecastingBayesian hierarchical modelscoherent forecastsprobabilistic forecastingtourism forecastinginformation sharing
0
0 comments X

The pith

A Bayesian model embeds hierarchical coherence and level priorities directly into forecast parameter estimation.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a Bayesian approach to hierarchical forecasting where the model itself shares information across levels and penalizes incoherence during estimation. This replaces the common two-step process of independent forecasts followed by reconciliation. The result is parameter estimates tuned to decision goals at specific levels and coherent by construction, eliminating separate covariance estimation for multi-step ahead predictions. Demonstrations on simulated data and Australian tourism data show gains in predictive accuracy.

Core claim

We propose a fully Bayesian hierarchical forecasting framework that shares information more effectively between and across levels than reconciliation alone. This yields parameter estimates that are focused towards the forecasting goals and capture the requirement for coherency, removing the need to estimate covariance matrices for multi-step forecasting horizons.

What carries the argument

The fully Bayesian hierarchical model specification that incorporates soft penalization of incoherence and focuses estimation on decision-relevant levels.

If this is right

  • Forecasts at all levels are coherent by construction without a separate reconciliation step.
  • Parameter estimates are optimized with respect to the hierarchical structure and specific decision goals.
  • No covariance matrices need to be estimated for multi-step forecasting horizons.
  • Information is shared more effectively between and across levels during estimation.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The approach could extend to supply-chain or energy-demand hierarchies where decision goals vary sharply by level.
  • Performance gains may increase in systems with stronger cross-level error dependencies than the tourism example.
  • Combining the framework with state-space or other dynamic Bayesian components offers a natural next step.

Load-bearing premise

The hierarchical structure and decision goals can be encoded into a Bayesian model specification such that soft penalization of incoherence and level-focused estimation produce superior parameter estimates without post-hoc reconciliation or covariance estimation.

What would settle it

A head-to-head test on the same simulated and Australian tourism datasets where post-hoc reconciled independent forecasts match or exceed the Bayesian model's predictive accuracy metrics would falsify the claimed advantage.

Figures

Figures reproduced from arXiv: 2606.23009 by Arwen Nugteren, Christopher Drovandi, Kerrie Mengersen, Mahdi Abolghasemi.

Figure 1
Figure 1. Figure 1: , with associated structural matrix: S =                    1 1 1 1 1 1 0 0 0 0 1 1 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1                    . Total A AA AB B BA BB [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: A figure illustrating all three aspects of our methodology – shared hierarchical prior, [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Nemenyi matrices for the overall accuracy metrics, averaged over all horizons. [PITH_FULL_IMAGE:figures/full_fig_p029_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Average RMSE over all horizons at each hierarchical level, split up by method. [PITH_FULL_IMAGE:figures/full_fig_p030_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Average CRPS over all horizons at each hierarchical level, split up by method. [PITH_FULL_IMAGE:figures/full_fig_p030_5.png] view at source ↗
read the original abstract

Decision-making in hierarchical systems requires probabilistic forecasts at all cross-sectional levels. Current hierarchical forecasting methods typically generate independent forecasts at each level and reconcile them post hoc to ensure coherence between upper and lower levels. Such post hoc corrections do not incorporate hierarchical structure or decision goals into the underlying parameter estimation. We propose a fully Bayesian hierarchical forecasting framework that shares information more effectively between and across levels than reconciliation alone. Our approach has the flexibility to softly penalise incoherence, subject to model specification, and to focus the global model and coherence update on hierarchical levels most relevant to decision outcomes. This yields parameter estimates that are focused towards the forecasting goals and capture the requirement for coherency, removing the need to estimate covariance matrices for multi-step forecasting horizons. We demonstrate improvements in predictive accuracy metrics on both simulated data and Australian domestic tourism forecasting.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript proposes a fully Bayesian hierarchical forecasting framework that encodes hierarchical structure and decision goals into parameter estimation via soft penalization of incoherence. This is claimed to share information across levels more effectively than post-hoc reconciliation, produce level-focused parameter estimates, enforce coherency through the model specification, and eliminate the need for covariance matrix estimation at multi-step horizons. Improvements in predictive accuracy are demonstrated on simulated data and Australian domestic tourism data.

Significance. If the soft penalization reliably produces coherent forecasts without post-hoc steps and the level-focusing improves accuracy, the framework would advance hierarchical forecasting by integrating constraints directly into Bayesian estimation rather than reconciliation. The Bayesian information-sharing aspect and avoidance of covariance estimation for multi-step horizons are potential strengths if substantiated.

major comments (2)
  1. [Abstract] Abstract: The central claim that the approach 'removes the need to estimate covariance matrices for multi-step forecasting horizons' and achieves coherency via soft penalization is load-bearing but unsupported by any model equations, likelihood/prior specifications, or penalty term details. Without these, it is impossible to verify whether the penalization enforces coherence or merely encourages it, leaving open the possibility that sampled forecasts still require implicit reconciliation.
  2. [Abstract] Abstract and demonstration sections: The reported improvements in predictive accuracy on simulated and tourism data lack any description of baselines, metrics (e.g., CRPS, RMSE), error analysis, or comparison to standard reconciliation methods. This prevents assessment of whether the claimed advantages over reconciliation are realized or if they reduce to data-specific fitting.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments. We address each major comment point by point below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim that the approach 'removes the need to estimate covariance matrices for multi-step forecasting horizons' and achieves coherency via soft penalization is load-bearing but unsupported by any model equations, likelihood/prior specifications, or penalty term details. Without these, it is impossible to verify whether the penalization enforces coherence or merely encourages it, leaving open the possibility that sampled forecasts still require implicit reconciliation.

    Authors: The abstract is a concise summary; the full model specification (hierarchical structure, likelihood, priors, and the incoherence penalization term added to the posterior) appears in Sections 2–3 of the manuscript. The soft penalization encourages coherence during parameter estimation, and the hierarchical Bayesian sampling produces level-consistent forecasts by construction, eliminating separate covariance estimation for multi-step horizons and post-hoc reconciliation. We agree the abstract could better signal these elements and will revise it to reference the key modeling choices. revision: partial

  2. Referee: [Abstract] Abstract and demonstration sections: The reported improvements in predictive accuracy on simulated and tourism data lack any description of baselines, metrics (e.g., CRPS, RMSE), error analysis, or comparison to standard reconciliation methods. This prevents assessment of whether the claimed advantages over reconciliation are realized or if they reduce to data-specific fitting.

    Authors: The abstract summarizes results at a high level. The manuscript's experimental sections detail the baselines (independent per-level forecasts plus standard reconciliation such as MinT), metrics (CRPS and RMSE), error analysis, and direct comparisons on both the simulated and Australian tourism data. We will revise the abstract to name the metrics and note the reconciliation comparisons. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper proposes a Bayesian hierarchical model that encodes structure and decision goals via soft penalization of incoherence in the model specification itself. The abstract and provided excerpts describe this as a modeling choice that produces coherent parameter estimates without post-hoc reconciliation or covariance estimation. No equations or steps are shown that reduce a claimed prediction to a fitted input by construction, nor is there self-definitional equivalence, load-bearing self-citation of an unverified uniqueness result, or renaming of known patterns. The derivation chain remains self-contained as a standard Bayesian specification exercise, with the central claim resting on the modeling approach rather than tautological reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Ledger entries are inferred from the abstract only; no specific free parameters, invented entities, or detailed axioms are provided in the available text.

axioms (2)
  • domain assumption Bayesian hierarchical models can share information across levels more effectively than independent forecasts followed by reconciliation.
    Core premise of the proposed framework.
  • domain assumption Soft penalization subject to model specification can enforce coherence while allowing level-specific focus.
    Key flexibility claimed for the approach.

pith-pipeline@v0.9.1-grok · 5677 in / 1264 out tokens · 43642 ms · 2026-06-26T07:43:23.818638+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

237 extracted references · 114 canonical work pages

  1. [1]

    Forecasting: Principles and Practice

    Hyndman, Robin John and George Athanasopoulos. Forecasting: Principles and Practice. 2021

    2021

  2. [2]

    2023 , issn =

    Forecast combinations: An over 50-year review , journal =. 2023 , issn =. doi:10.1016/j.ijforecast.2022.11.005 , author =

    work page doi:10.1016/j.ijforecast.2022.11.005 2023

  3. [3]

    Probabilistic forecast reconciliation with applications to wind power and electric load , volume =

    Jeon, Jooyoung and Panagiotelis, Anastasios and Petropoulos, Fotios , year =. Probabilistic forecast reconciliation with applications to wind power and electric load , volume =. European Journal of Operational Research , doi =

  4. [4]

    2014 , issn =

    Global Energy Forecasting Competition 2012 , journal =. 2014 , issn =. doi:10.1016/j.ijforecast.2013.07.001 , url =

    work page doi:10.1016/j.ijforecast.2013.07.001 2012

  5. [5]

    Energy Forecasting: A Review and Outlook

    Tao Hong and Pierre Pinson and Yi Wang and Rafal Weron and Dazhi Yang and Hamidreza Zareipour. Energy Forecasting: A Review and Outlook. IEEE Open Access Journal of Power and Energy. 2020. doi:10.1109/OAJPE.2020.3029979

    work page doi:10.1109/oajpe.2020.3029979 2020

  6. [6]

    Martin and David T

    Ruben Loaiza-Maya and Gael M. Martin and David T. Frazier , year=. Focused. Journal of Applied Econometrics

  7. [7]

    2023 , eprint=

    Bayesian Predictive Decision Synthesis , author=. 2023 , eprint=

    2023

  8. [8]

    2016 , issn =

    Probabilistic electric load forecasting: A tutorial review , journal =. 2016 , issn =. doi:10.1016/j.ijforecast.2015.11.011 , url =

    work page doi:10.1016/j.ijforecast.2015.11.011 2016

  9. [9]

    Proceedings of the 34th International Conference on Machine Learning , pages =

    Coherent Probabilistic Forecasts for Hierarchical Time Series , author =. Proceedings of the 34th International Conference on Machine Learning , pages =. 2017 , editor =

    2017

  10. [10]

    Journal of the American Statistical Association , volume =

    Federico Bassetti and Roberto Casarin and Francesco Ravazzolo , title =. Journal of the American Statistical Association , volume =. 2018 , publisher =

    2018

  11. [13]

    2021 , issn =

    Forecast reconciliation: A geometric view with new insights on bias correction , journal =. 2021 , issn =. doi:10.1016/j.ijforecast.2020.06.004 , url =

    work page doi:10.1016/j.ijforecast.2020.06.004 2021

  12. [14]

    2024 , issn =

    Decision-focused learning under decision dependent uncertainty for power systems with price-responsive demand , journal =. 2024 , issn =. doi:10.1016/j.epsr.2024.110665 , url =

    work page doi:10.1016/j.epsr.2024.110665 2024

  13. [15]

    2025 , eprint=

    Value-oriented forecast reconciliation for renewables in electricity markets , author=. 2025 , eprint=

    2025

  14. [18]

    Wickramasuriya and George Athanasopoulos and Rob J

    Shanika L. Wickramasuriya and George Athanasopoulos and Rob J. Hyndman , title =. Journal of the American Statistical Association , volume =. 2019 , publisher =

    2019

  15. [19]

    Hierarchical Probabilistic Forecasting of Electricity Demand With Smart Meter Data , journal =

    Souhaib. Hierarchical Probabilistic Forecasting of Electricity Demand With Smart Meter Data , journal =. 2021 , publisher =

    2021

  16. [20]

    Multisensor Data Fusion

    Durrant-Whyte, Hugh and Henderson, Thomas C. Multisensor Data Fusion. Springer Handbook of Robotics. 2008. doi:10.1007/978-3-540-30301-5_26

    work page doi:10.1007/978-3-540-30301-5_26 2008

  17. [21]

    and Grushka-Cockayne, Yael and Winkler, Robert L

    Lichtendahl, Kenneth C. and Grushka-Cockayne, Yael and Winkler, Robert L. , title =. Management Science , volume =. 2013 , doi =

    2013

  18. [22]

    Ali Hassani, Steven Walton, Jiachen Li, Shen Li, and Humphrey Shi

    Tilmann Gneiting and Adrian E Raftery , title =. Journal of the American Statistical Association , volume =. 2007 , publisher =. doi:10.1198/016214506000001437 , eprint =

    work page doi:10.1198/016214506000001437 2007

  19. [23]

    2016 , issn =

    The forecast combination puzzle: A simple theoretical explanation , journal =. 2016 , issn =. doi:10.1016/j.ijforecast.2015.12.005 , url =

    work page doi:10.1016/j.ijforecast.2015.12.005 2016

  20. [24]

    International Journal of Forecasting , volume =

    Some theoretical results on forecast combinations , author =. International Journal of Forecasting , volume =. 2018 , issn =. doi:10.1016/j.ijforecast.2017.08.005 , url =

    work page doi:10.1016/j.ijforecast.2017.08.005 2018

  21. [25]

    On Normality Assumption in Residual Simulation for Probabilistic Load Forecasting , year=

    Xie, Jingrui and Hong, Tao and Laing, Thomas and Kang, Chongqing , journal=. On Normality Assumption in Residual Simulation for Probabilistic Load Forecasting , year=

  22. [27]

    2020 , issn =

    Cross-temporal aggregation: Improving the forecast accuracy of hierarchical electricity consumption , journal =. 2020 , issn =. doi:10.1016/j.apenergy.2019.114339 , author =

    work page doi:10.1016/j.apenergy.2019.114339 2020

  23. [28]

    2021 , eprint=

    Simultaneously Reconciled Quantile Forecasting of Hierarchically Related Time Series , author=. 2021 , eprint=

    2021

  24. [29]

    Proceedings of the 38th International Conference on Machine Learning , pages =

    End-to-End Learning of Coherent Probabilistic Forecasts for Hierarchical Time Series , author =. Proceedings of the 38th International Conference on Machine Learning , pages =. 2021 , editor =

    2021

  25. [30]

    2024 , issn =

    Machine learning applications in hierarchical time series forecasting: Investigating the impact of promotions , journal =. 2024 , issn =. doi:10.1016/j.ijforecast.2022.07.004 , url =

    work page doi:10.1016/j.ijforecast.2022.07.004 2024

  26. [31]

    and White, Ethan P

    Simonis, Juniper L. and White, Ethan P. and Ernest, S. K. Morgan , title =. Ecology , volume =. doi:10.1002/ecy.3431 , year =

    work page doi:10.1002/ecy.3431

  27. [32]

    and Hou, Aiqin , booktitle=

    Zhang, Minjuan and Wu, Chase Q. and Hou, Aiqin , booktitle=. Ensemble Learning Models for Large-Scale Time Series Forecasting in Supply Chain , year=

  28. [33]

    Journal of the American Statistical Association , volume =

    Sam Allen and Jonathan Koh and Johan Segers and Johanna Ziegel and , title =. Journal of the American Statistical Association , volume =. 2025 , publisher =

    2025

  29. [34]

    and Gianfreda, Angelica and Kermer, Stefan , title =

    Bunn, Derek W. and Gianfreda, Angelica and Kermer, Stefan , title =. Energies , volume =. 2018 , number =

    2018

  30. [35]

    2019 , issn =

    Interpreting the skill score form of forecast performance metrics , journal =. 2019 , issn =. doi:10.1016/j.ijforecast.2018.11.010 , author =

    work page doi:10.1016/j.ijforecast.2018.11.010 2019

  31. [36]

    2021 , issn =

    An introduction to multivariate probabilistic forecast evaluation , journal =. 2021 , issn =. doi:10.1016/j.egyai.2021.100058 , url =

    work page doi:10.1016/j.egyai.2021.100058 2021

  32. [37]

    Journal of Statistical Modeling and Analytics , volume =

    Nornadiah Mohd Razali and Yap Bee Wah , title=. Journal of Statistical Modeling and Analytics , volume =. 2011 , url=

    2011

  33. [38]

    Roulston and Leonard A

    Mark S. Roulston and Leonard A. Smith. Evaluating Probabilistic Forecasts Using Information Theory. Monthly Weather Review. 2002. doi:10.1175/1520-0493(2002)130<1653:EPFUIT>2.0.CO;2

    work page doi:10.1175/1520-0493(2002)130 2002

  34. [39]

    Siegert, Stefan and Ferro, Christopher A. T. and Stephenson, David B. and Leutbecher, Martin , title =. Quarterly Journal of the Royal Meteorological Society , volume =. 2019 , doi =

    2019

  35. [40]

    Joint Bottom-up Method for Probabilistic Forecasting of Hierarchical Time Series , journal =

    Bertani, Nicol\`. Joint Bottom-up Method for Probabilistic Forecasting of Hierarchical Time Series , journal =. 2025 , doi =

    2025

  36. [43]

    Tilmann Gneiting and Adrian E

    Gneiting, Tilmann and Katzfuss, Matthias. Probabilistic Forecasting. Annual Review of Statistics and Its Application. 2014. doi:10.1146/annurev-statistics-062713-085831

    work page doi:10.1146/annurev-statistics-062713-085831 2014

  37. [44]

    Box, G. E. P. and Cox, D. R. , title =. Journal of the Royal Statistical Society: Series B (Methodological) , volume =. 1964 , month =. doi:10.1111/j.2517-6161.1964.tb00553.x , eprint =

    work page doi:10.1111/j.2517-6161.1964.tb00553.x 1964

  38. [45]

    International Journal of Forecasting , volume =

    Exponential smoothing: The state of the art—Part. International Journal of Forecasting , volume =. 2006 , issn =. doi:10.1016/j.ijforecast.2006.03.005 , author =

    work page doi:10.1016/j.ijforecast.2006.03.005 2006

  39. [46]

    Stéphane Vannitsem and John Bjørnar Bremnes and Jonathan Demaeyer and Gavin R. Evans and Jonathan Flowerdew and Stephan Hemri and Sebastian Lerch and Nigel Roberts and Susanne Theis and Aitor Atencia and Zied Ben Bouallègue and Jonas Bhend and Markus Dabernig and Lesley De Cruz and Leila Hieta and Olivier Mestre and Lionel Moret and Iris Odak Plenković an...

    work page doi:10.1175/bams-d-19-0308.1 2021

  40. [47]

    and Bauer, Peter and Adams, Samantha , publisher =

    Dueben, Peter D. and Bauer, Peter and Adams, Samantha , publisher =. Deep Learning to Improve Weather Predictions , booktitle =. doi:10.1002/9781119646181.ch14 , eprint =

    work page doi:10.1002/9781119646181.ch14

  41. [48]

    Journal of the American Statistical Association , volume =

    Tilmann Gneiting , title =. Journal of the American Statistical Association , volume =. 2011 , publisher =. doi:10.1198/jasa.2011.r10138 , eprint =

    work page doi:10.1198/jasa.2011.r10138 2011

  42. [49]

    Control Engineering Practice , volume =

    Quantile forecast of renewable energy generation based on Indicator Gradient Descent and deep residual. Control Engineering Practice , volume =. 2021 , issn =. doi:10.1016/j.conengprac.2021.104863 , url =

    work page doi:10.1016/j.conengprac.2021.104863 2021

  43. [50]

    2020 , issn =

    Probabilistic State Forecasting and Optimal Voltage Control in Distribution Grids under Uncertainty , journal =. 2020 , issn =. doi:10.1016/j.epsr.2020.106562 , url =

    work page doi:10.1016/j.epsr.2020.106562 2020

  44. [51]

    Econometrica , volume =

    Chernozhukov, Victor and Fernández-Val, Iván and Galichon, Alfred , title =. Econometrica , volume =. doi:10.3982/ECTA7880 , eprint =

    work page doi:10.3982/ecta7880

  45. [52]

    and Pillai, Natesh and Park, Ju-Hyun , number =

    Dunson, David B. and Pillai, Natesh and Park, Ju-Hyun , title =. Journal of the Royal Statistical Society Series B: Statistical Methodology , volume =. 2007 , month =. doi:10.1111/j.1467-9868.2007.00582.x , eprint =

    work page doi:10.1111/j.1467-9868.2007.00582.x 2007

  46. [53]

    2019 , issn =

    Bayesian deep learning based method for probabilistic forecast of day-ahead electricity prices , journal =. 2019 , issn =. doi:10.1016/j.apenergy.2019.05.068 , url =

    work page doi:10.1016/j.apenergy.2019.05.068 2019

  47. [54]

    Chen, Xia and Dong, Zhao Yang and Meng, Ke and Xu, Yan and Wong, Kit Po and Ngan, H. W. , journal=. Electricity Price Forecasting With Extreme Learning Machine and Bootstrapping , year=

  48. [55]

    and Raftery, A

    Alkema, L. and Raftery, A. E. and Gerland, P. and Clark, S. J. and Pelletier, F. and Buettner, T. and Heilig, G. K. , title =. Demography , volume =. 2011 , doi =

    2011

  49. [56]

    Raftery and Samuel J

    Leontine Alkema and Adrian E. Raftery and Samuel J. Clark , title =. Annals of Applied Statistics , volume =. 2007 , doi =

    2007

  50. [57]

    Monthly Weather Review , author =

    Adrian E. Raftery and Tilmann Gneiting and Fadoua Balabdaoui and Michael Polakowski. Using Bayesian Model Averaging to Calibrate Forecast Ensembles. Monthly Weather Review. 2005. doi:10.1175/MWR2906.1

    work page doi:10.1175/mwr2906.1 2005

  51. [58]

    2010 , journal =

    Ranjan, Roopesh and Gneiting, Tilmann , title =. Journal of the Royal Statistical Society: Series B (Statistical Methodology) , volume =. doi:10.1111/j.1467-9868.2009.00726.x , eprint =

    work page doi:10.1111/j.1467-9868.2009.00726.x 2009

  52. [59]

    2008 , issn =

    Bayesian Model Averaging and exchange rate forecasts , journal =. 2008 , issn =. doi:10.1016/j.jeconom.2008.08.012 , url =

    work page doi:10.1016/j.jeconom.2008.08.012 2008

  53. [60]

    International Journal of Forecasting , volume =

    Comparing and evaluating. International Journal of Forecasting , volume =. 2010 , note =. doi:10.1016/j.ijforecast.2009.10.007 , url =

    work page doi:10.1016/j.ijforecast.2009.10.007 2010

  54. [61]

    Taylor and , title =

    Jooyoung Jeon and James W. Taylor and , title =. Journal of the American Statistical Association , volume =. 2012 , publisher =. doi:10.1080/01621459.2011.643745 , eprint =

    work page doi:10.1080/01621459.2011.643745 2012

  55. [62]

    2016 , issn =

    Probabilistic energy forecasting: Global Energy Forecasting Competition 2014 and beyond , journal =. 2016 , issn =. doi:10.1016/j.ijforecast.2016.02.001 , url =

    work page doi:10.1016/j.ijforecast.2016.02.001 2014

  56. [63]

    Bracale, Antonio and Caramia, Pierluigi and Carpinelli, Guido and Di Fazio, Anna Rita and Varilone, Pietro , journal=. A. 2013 , volume=

    2013

  57. [64]

    and Ohmi, M

    Mori, H. and Ohmi, M. , booktitle=. Probabilistic short-term load forecasting with. 2005 , pages=

    2005

  58. [65]

    2015 , issn =

    Examining performance of aggregation algorithms for neural network-based electricity demand forecasting , journal =. 2015 , issn =. doi:10.1016/j.ijepes.2014.08.025 , url =

    work page doi:10.1016/j.ijepes.2014.08.025 2015

  59. [66]

    Journal of Building Engineering , volume =

    Probabilistic electrical load forecasting for buildings using. Journal of Building Engineering , volume =. 2022 , issn =. doi:10.1016/j.jobe.2021.103853 , url =

    work page doi:10.1016/j.jobe.2021.103853 2022

  60. [67]

    Energy Consumption Forecasting Using a Stacked Nonparametric Bayesian Approach

    Weeraddana, Dilusha and Khoa, Nguyen Lu Dang and O'Neil, Lachlan and Wang, Weihong and Cai, Chen. Energy Consumption Forecasting Using a Stacked Nonparametric Bayesian Approach. Machine Learning and Knowledge Discovery in Databases. Applied Data Science and Demo Track. 2021

    2021

  61. [68]

    and Qiu, G

    Fan, H. and Qiu, G. and Guo, X. and Zhen, Z. and Wang, F. and Mi, Z. and Wang, G. , booktitle=. WIND SPEED PROBABILISTIC FORECASTING BASED ON THE CORRELATION BETWEEN WIND PATTERNS AND PREDICTION ERRORS , year=

  62. [69]

    A Nonparametric

    Sun, Weigao and Zamani, Mohsen , booktitle=. A Nonparametric. 2019 , pages=

    2019

  63. [70]

    2022 , howpublished =

    Notation for forecast reconciliation , author =. 2022 , howpublished =

    2022

  64. [71]

    Wickramasuriya , title =

    Shanika L. Wickramasuriya , title =. Journal of Business & Economic Statistics , volume =. 2024 , publisher =

    2024

  65. [72]

    2025 , issn =

    Optimal forecast reconciliation with time series selection , journal =. 2025 , issn =. doi:10.1016/j.ejor.2024.12.004 , url =

    work page doi:10.1016/j.ejor.2024.12.004 2025

  66. [73]

    , title =

    Winkler, Robert L. , title =. Journal of the American Statistical Association , volume =. 1972 , doi =

    1972

  67. [74]

    Electronic Journal of Statistics , year =

    Sebastian Arnold and Eva-Maria Walz and Johanna Ziegel and Tilmann Gneiting , title =. Electronic Journal of Statistics , year =

  68. [75]

    Hyndman and Evangelos Spiliotis and Christoph Bergmeir , title =

    Mahdi Abolghasemi and Rob J. Hyndman and Evangelos Spiliotis and Christoph Bergmeir , title =. Machine learning , volume =. 2022 , doi =

    2022

  69. [76]

    Brehmer and Tilmann Gneiting , title =

    Jonas R. Brehmer and Tilmann Gneiting , title =. Bernoulli , number =. 2021 , doi =

    2021

  70. [77]

    and Gneiting, T , title =

    Brehmer, J.R. and Gneiting, T , title =. Annals of the Institute of Statistical Mathematics , year =

  71. [78]

    2019 , issn =

    Global energy forecasting competition 2017: Hierarchical probabilistic load forecasting , journal =. 2019 , issn =. doi:10.1016/j.ijforecast.2019.02.006 , url =

    work page doi:10.1016/j.ijforecast.2019.02.006 2017

  72. [79]

    Data visualization and forecast combination for probabilistic load forecasting in GEFCom2017 final match , journal =

    Julian. Data visualization and forecast combination for probabilistic load forecasting in GEFCom2017 final match , journal =. 2019 , issn =. doi:10.1016/j.ijforecast.2019.02.004 , url =

    work page doi:10.1016/j.ijforecast.2019.02.004 2019

  73. [80]

    2019 , issn =

    Data preprocessing and quantile regression for probabilistic load forecasting in the GEFCom2017 final match , journal =. 2019 , issn =. doi:10.1016/j.ijforecast.2019.02.005 , url =

    work page doi:10.1016/j.ijforecast.2019.02.005 2019

  74. [81]

    2025 , eprint=

    Informed Forecasting: Leveraging Auxiliary Knowledge to Boost LLM Performance on Time Series Forecasting , author=. 2025 , eprint=

    2025

  75. [82]

    2025 , eprint=

    Modeling the uncertainty on the covariance matrix for probabilistic forecast reconciliation , author=. 2025 , eprint=

    2025

  76. [83]

    2025 , eprint=

    Geometry of nonlinear forecast reconciliation , author=. 2025 , eprint=

    2025

  77. [84]

    and Lerch, Sebastian and Phipps, Kaleb and Schienle, Melanie

    Gneiting, Tilmann and Wolffram, Daniel and Resin, Johannes and Kraus, Kristof and Bracher, Johannes and Dimitriadis, Timo and Hagenmeyer, Veit and Jordan, Alexander I. and Lerch, Sebastian and Phipps, Kaleb and Schienle, Melanie. Model Diagnostics and Forecast Evaluation for Quantiles. Annual Review of Statistics and Its Application. 2023. doi:10.1146/ann...

    work page doi:10.1146/annurev-statistics-032921-020240 2023

  78. [85]

    2025 , issn =

    Bayesian hierarchical probabilistic forecasting of intraday electricity prices , journal =. 2025 , issn =. doi:10.1016/j.apenergy.2024.124975 , url =

    work page doi:10.1016/j.apenergy.2024.124975 2025

  79. [86]

    Julie Novak and Scott McGarvie and Beatriz Etchegaray Garcia , year=. A. 1711.04738 , archivePrefix=

    Pith/arXiv arXiv

  80. [87]

    2025 , eprint=

    The cost of ensembling: is it always worth combining? , author=. 2025 , eprint=

    2025

Showing first 80 references.