Metraipolis › archive › dc_1b38232daaf2
Global tree forecasters collapse at the hierarchical aggregate: a five-panel failure characterization — arXiv:2609.27912
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.27912
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T10:57:24.911Z
- SHA-256 of the archived text
44b4c368c7140e52555ca4d8e1ea6d9fae12537b8c01a077cceecf8e82c41e1a- Stable URL of this page
https://metraipolis.com/documento/dc_1b38232daaf2
How to cite
Global tree forecasters collapse at the hierarchical aggregate: a five-panel failure characterization — arXiv:2609.27912 — arXiv (en), https://arxiv.org/abs/2609.27912 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1b38232daaf2 (Metraipolis archive dc_1b38232daaf2, sha256 44b4c368c7140e52555ca4d8e1ea6d9fae12537b8c01a077cceecf8e82c41e1a)
BibTeX
@misc{dc1b38232daaf2,
title = {Global tree forecasters collapse at the hierarchical aggregate: a five-panel failure characterization — arXiv:2609.27912},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.27912},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1b38232daaf2; sha256 44b4c368c7140e52555ca4d8e1ea6d9fae12537b8c01a077cceecf8e82c41e1a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.27912. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Global forecasting models pool many series and learn one shared function. Gradient-boosted trees are their most common form. We measure a failure of this design that has not, to our knowledge, been documented. Train a global tree on the individual series of a hierarchy, then ask it for the hierarchical aggregate. The aggregate sits far outside the model's training range, and the forecast collapses. The model under-predicts the total by 30-50x in our production deployment, and by up to 496x in a public M5 reconstruction. The mechanism is known: beyond its training range, a tree predicts a constant. It surfaces at the aggregate because the total dwarfs every training series. The cure is not new. Per-series scaling, the preprocessing step that Montero-Manso and Hyndman (2021) recommend, prevents the collapse. So do a weighted aggregate-level training row and seasonal differencing. Our contribution is the characterization. The collapse reproduces on five panels: a production business-to-business marketplace, a synthetic hierarchy, M5, Australian Tourism, and a public business-buyer panel. It holds on three tree libraries, is invariant across training seeds, and is statistically significant. Its onset is immediate and tracks a simple support bound: a scale gap of only 1.15x already costs a third of the total. No standard configuration change prevents it: pooling every hierarchy level into training fails at scale, and the one knob that fits linear models in the leaves softens it without curing it. Rolling the forecasts forward recursively separates the cures: the aggregate-row cure re-collapses, per-series scaling degrades but stays low, and only seasonal differencing keeps its one-step accuracy unchanged. We close with a three-step procedure for diagnosing and preventing the failure in deployed systems. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-24 in cs.LG). Categorie arXiv: Machine Learning (cs.LG) ; Econometrics (econ.EM); Applications (stat.AP). Pagina del lavoro: https://arxiv.org/abs/2609.27912 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-08T10:06:16.755Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-24?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).