IDEAS home Printed from https://ideas.repec.org/p/arx/papers/2609.23378.html

Leaky-integrator reconstruction: taming error accumulation in recursive differenced time-series forecasting

Author

Listed:
  • Zijiang Yang

Abstract

Recursive differenced forecasting, the standard remedy for non-stationarity, predicts one-step changes and integrates them by cumulative summation. We show that this reconstruction is a discrete integrator with a pole on the unit circle, so the biased increment errors of a learned nonlinear model are summed without bound and the rollout diverges: at 336 steps its normalised MAE reaches 1.6-3.8 for every neural architecture tested, against 0.80 for a stable linear recursion. We then introduce leaky-integrator reconstruction, a training-free fix that moves the pole inside the unit circle with H(z) = 1/(1 - gamma z^-1), gamma

Suggested Citation

  • Zijiang Yang, 2026. "Leaky-integrator reconstruction: taming error accumulation in recursive differenced time-series forecasting," Papers 2609.23378, arXiv.org, revised Sep 2026.
    • Handle: RePEc:arx:papers:2609.23378
    as

    Download full text from publisher

    File URL: https://arxiv.org/pdf/2609.23378
    File Function: Latest version
    Download Restriction: no
    ---><---

    References listed on IDEAS

    as
    1. Salinas, David & Flunkert, Valentin & Gasthaus, Jan & Januschowski, Tim, 2020. "DeepAR: Probabilistic forecasting with autoregressive recurrent networks," International Journal of Forecasting, Elsevier, vol. 36(3), pages 1181-1191.
    2. Marcellino, Massimiliano & Stock, James H. & Watson, Mark W., 2006. "A comparison of direct and iterated multistep AR methods for forecasting macroeconomic time series," Journal of Econometrics, Elsevier, vol. 135(1-2), pages 499-526.
    Full references (including those not matched with items on IDEAS)

    Most related items

    These are the items that most often cite the same works as this one and are cited by the same works as this one.
    1. Md Rezwanul Islam & Wael Mohammed, 2026. "Global tree forecasters collapse at the hierarchical aggregate: a five-panel failure characterization," Papers 2609.27912, arXiv.org.
    2. Anna Almosova & Niek Andresen, 2023. "Nonlinear inflation forecasting with recurrent neural networks," Journal of Forecasting, John Wiley & Sons, Ltd., vol. 42(2), pages 240-259, March.
    3. Zhen Zeng & Rachneet Kaur & Suchetha Siddagangappa & Saba Rahimi & Tucker Balch & Manuela Veloso, 2023. "Financial Time Series Forecasting using CNN and Transformer," Papers 2304.04912, arXiv.org.
    4. Likai Chen & Weining Wang, 2026. "From Vector Autoregressions to AI-based Time Series Forecasting: A Review," Bristol Economics Discussion Papers 26/838, School of Economics, University of Bristol, UK.
    5. Likai Chen & Weining Wang, 2026. "From Vector Autoregressions to AI-based Time Series Forecasting: A Review," Papers 2607.14279, arXiv.org.
    6. Hauzenberger, Niko & Huber, Florian & Klieber, Karin & Marcellino, Massimiliano, 2025. "Bayesian neural networks for macroeconomic analysis," Journal of Econometrics, Elsevier, vol. 249(PC).
    7. Eric Ghysels & Leonardo Iania & Jonas Striaukas, 2018. "Quantile-based Inflation Risk Models," Working Paper Research 349, National Bank of Belgium.
    8. Galvão, Ana Beatriz, 2013. "Changes in predictive ability with mixed frequency data," International Journal of Forecasting, Elsevier, vol. 29(3), pages 395-410.
    9. Eliana Gonz�lez & Luis F. Melo & Viviana Monroy & Brayan Rojas, 2009. "A Dynamic Factor Model For The Colombian Inflation," Borradores de Economia 5273, Banco de la Republica.
    10. Andrea Carriero & Todd E. Clark & Massimiliano Marcellino, 2015. "Realtime nowcasting with a Bayesian mixed frequency model with stochastic volatility," Journal of the Royal Statistical Society Series A, Royal Statistical Society, vol. 178(4), pages 837-862, October.
    11. Nason, Guy P. & Palasciano, Henry Antonio, 2026. "Forecasting UK consumer price inflation with RaGNAR: Random generalised network autoregressive processes," International Journal of Forecasting, Elsevier, vol. 42(1), pages 181-202.
    12. Ali Atiah Alzahrani, 2026. "Verify Claims, Not Scores: Evidence-Based Verification of Modular Agents," Papers 2610.01348, arXiv.org.
    13. Michael W. McCracken & Michael T. Owyang & Tatevik Sekhposyan, 2021. "Real-Time Forecasting and Scenario Analysis Using a Large Mixed-Frequency Bayesian VAR," International Journal of Central Banking, International Journal of Central Banking, vol. 17(71), pages 1-41, December.
    14. Philippe Goulet Coulombe & Maxime Leroux & Dalibor Stevanovic & Stéphane Surprenant, 2022. "How is machine learning useful for macroeconomic forecasting?," Journal of Applied Econometrics, John Wiley & Sons, Ltd., vol. 37(5), pages 920-964, August.
    15. Ferrari, Davide & Ravazzolo, Francesco & Vespignani, Joaquin, 2021. "Forecasting energy commodity prices: A large global dataset sparse approach," Energy Economics, Elsevier, vol. 98(C).
    16. Kelly Trinh & Bo Zhang & Chenghan Hou, 2025. "Macroeconomic real‐time forecasts of univariate models with flexible error structures," Journal of Forecasting, John Wiley & Sons, Ltd., vol. 44(1), pages 59-78, January.
    17. Zeng, Huanze & Shi, Chenlu & Fang, Haoyu & Wu, Binrong, 2025. "Interpretable multivariate wind speed forecasting using sliding masked window-based decomposition and deep autoregressive networks," Energy, Elsevier, vol. 341(C).
    18. Stefano Puddu, 2013. "Real Sector and Banking System: Real and Feedback Effects. A Non-Linear VAR Approach," IRENE Working Papers 13-01, IRENE Institute of Economic Research.
    19. Ivan S. Maksymov, 2023. "Analogue and Physical Reservoir Computing Using Water Waves: Applications in Power Engineering and Beyond," Energies, MDPI, vol. 16(14), pages 1-26, July.
    20. Nonejad, Nima, 2022. "An interesting finding about the ability of geopolitical risk to forecast aggregate equity return volatility out-of-sample," Finance Research Letters, Elsevier, vol. 47(PB).

    More about this item

    NEP fields

    This paper has been announced in the following NEP Reports:

    Statistics

    Access and download statistics

    Corrections

    All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:arx:papers:2609.23378. See general information about how to correct material in RePEc.

    If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.

    If CitEc recognized a bibliographic reference but did not link an item in RePEc to it, you can help with this form .

    If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.

    For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: arXiv administrators (email available below). General contact details of provider: https://arxiv.org/ .

    Please note that corrections may take a couple of weeks to filter through the various RePEc services.

    IDEAS is a RePEc service. RePEc uses bibliographic data supplied by the respective publishers.