EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

RobustiPy: An efficient next-generation multiversal library with model selection, averaging, resampling, and explainable AI

Daniel Valdenegro, Jiani Yan, Duiyi Dai, Charles Rahal

Patterns · 2026-08 · 7 (8) · 101609

DOI: 10.1016/j.patter.2026.101609 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Patterns, 2026-08. Verfasst von Daniel Valdenegro und 3 weiteren Autor:innen. Beteiligte Einrichtung: Leverhulme Centre for Demographic Science, University of Oxford, Oxford, UK. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 76 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: The Leverhulme Trust, University of Oxford Nuffield College, UKRI ESRC.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Scientific inference is undermined by the vast but rarely explored "multiverse" of defensible modeling choices, which can generate results as variable as the phenomena under study. We introduce RobustiPy, an open-source Python library that systematizes multiverse analysis and model uncertainty quantification at scale. RobustiPy unifies bootstrap-based inference, combinatorial specification search, model selection and averaging, joint inference routines, and explainable AI methods within a modular, reproducible framework. Beyond exhaustive specification curves, it supports rigorous out-of-sample validation and provides feature-level explanations for the full-specification predictive model. We demonstrate utility across five simulations and ten high-profile replications spanning economics, sociology, psychology, and medicine, including a re-analysis of widely cited findings with documented discrepancies. Benchmarking on ∼672 million simulated regressions shows that RobustiPy delivers state-of-the-art computational efficiency while expanding transparency. By standardizing and accelerating methods for robustness, RobustiPy transforms how researchers interrogate sensitivity across the analytical multiverse, offering a foundation for reproducible and interpretable computational science.

Schlagwörter

Stichwörter der Autor:innen

Bayesian model averagingbootstrap resamplingexplainable AI/SHAPjoint inferencemodel selectionmodel uncertainty quantificationmultiverse analysisout-of-sample validationrobustness analysisspecification curve analysis

Autor:innen und Institutionen

  • Leverhulme Centre for Demographic Science, University of Oxford, Oxford, UK.
  • ESRC Centre for Care, University of Oxford, Oxford, UK.
  • Nuffield College, University of Oxford, Oxford, UK.
  • Department of Sociology, University of Oxford, Oxford, UK.
  • Wolfson College, University of Oxford, Oxford, UK.

Förderung

  • The Leverhulme Trust
  • University of Oxford Nuffield College
  • UKRI ESRC

Zitieren

APA

Valdenegro, D., Yan, J., Dai, D., & Rahal, C. (2026). RobustiPy: An efficient next-generation multiversal library with model selection, averaging, resampling, and explainable AI. Patterns, 7(8), 101609. https://doi.org/10.1016/j.patter.2026.101609

Vancouver

Valdenegro D, Yan J, Dai D, Rahal C. RobustiPy: An efficient next-generation multiversal library with model selection, averaging, resampling, and explainable AI. Patterns. 2026;7(8):101609. doi:10.1016/j.patter.2026.101609

BibTeX
@article{Valdenegro2026,
  author = {Valdenegro, Daniel and Yan, Jiani and Dai, Duiyi and Rahal, Charles},
  title = {RobustiPy: An efficient next-generation multiversal library with model selection, averaging, resampling, and explainable AI},
  journal = {Patterns},
  year = {2026},
  volume = {7},
  number = {8},
  pages = {101609},
  doi = {10.1016/j.patter.2026.101609}
}
Kostenlos starten →