Proximal R2 Priors for Structured Variance Allocation

Abstract

Scientific domain knowledge in regression often concerns structural relationships among model terms, which are typically encoded as parameter constraints. Since such relationships rarely hold exactly in practice, prior distributions should encourage them while at the same time retaining support for scientifically plausible departures. Accounting for these relationships becomes more pressing in modern regression, where high-dimensionality is usually present and regularisation is essential. R² decomposition priors provide an interpretable foundation for regularisation by placing a prior on total explained variance and distributing the implied signal across model terms. We introduce the Proximal R2 prior, which uses proximal relaxations to encode these relationships as soft constraints on the variance allocation. Our formulation yields differentiable posteriors compatible with standard gradient based samplers. Theoretically, we prove that our prior preserves the characteristic marginal properties of global-local shrinkage priors. Moreover, we establish posterior consistency and contraction rates in growing dimensions, showing that the inclusion of structural relationships and proximal methods do not alter the shrinkage properties. Notably, these results demonstrate that the structural constraints in the model do not hinder posterior recovery. Simulation studies and real-data applications to interaction and generalised additive models demonstrate that our prior improves parameter recovery, out-of-sample prediction, and sampling efficiency over existing constrained and unconstrained shrinkage priors, even under structural misspecification.

R2-GP: Variance-Calibrated Priors for Additive Gaussian Processes
Javier Enrique Aguilar, Soham Mukherjee, David Kohns
arXiv19 Oct 2026PreprintCC BY 4.0Full textStatisticsarXiv
Dirichlet-Tree Allocation Priors for Bayesian Causal Forests
Javier Enrique Aguilar, Daniel Klippert, Yusufhan Balci +1
arXiv19 Oct 2026PreprintCC BY 4.0Full textStatisticsarXiv
A tutorial on selecting, translating, and reporting variance priors in Bayesian multilevel models.
Liu Liu
Behavior research methods19 Oct 2026Peer-reviewedCC BYFull textStatisticsEurope PMC
Bayesian Generalized Network Autoregressive Model with Structured Shrinkage and Persistence Priors
Seongmin Kim, Kyusoon Kim
arXiv26 Oct 2026PreprintarXiv licenceFull textStatisticsarXiv
Estimation-Robust Minimum-Variance Portfolio Allocation Under Factor-Loading Uncertainty
Leonardo P. Carvalho, Jonathan M. Palma
Mathematics15 Oct 2026Peer-reviewedCC BY 4.0General mathematicsCrossref

Cite

BibTeX RIS

BibTeX

@misc{aguilar2026proximal,
  title = {{Proximal R2 Priors for Structured Variance Allocation}},
  author = {Javier Enrique Aguilar and David Kohns and Eric Yanchenko and Shonosuke Sugasawa and Kaoru Irie and Aki Vehtari and Paul-Christian Bürkner},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.38879},
  eprint = {2610.38879},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.38879},
  note = {CC BY 4.0},
}

RIS

TY  - UNPB
TI  - Proximal R2 Priors for Structured Variance Allocation
AU  - Javier Enrique Aguilar
AU  - David Kohns
AU  - Eric Yanchenko
AU  - Shonosuke Sugasawa
AU  - Kaoru Irie
AU  - Aki Vehtari
AU  - Paul-Christian Bürkner
PY  - 2026
DA  - 2026/10/28
PB  - arXiv
DO  - 10.48550/arxiv.2610.38879
UR  - https://arxiv.org/abs/2610.38879
AB  - Scientific domain knowledge in regression often concerns structural relationships among model terms, which are typically encoded as parameter constraints. Since such relationships rarely hold exactly in practice, prior distributions should encourage them while at the same time retaining support for scientifically plausible departures. Accounting for these relationships becomes more pressing in modern regression, where high-dimensionality is usually present and regularisation is essential. R^2 decomposition priors provide an interpretable foundation for regularisation by placing a prior on total explained variance and distributing the implied signal across model terms. We introduce the Proximal R2 prior, which uses proximal relaxations to encode these relationships as soft constraints on the variance allocation. Our formulation yields differentiable posteriors compatible with standard gradient based samplers. Theoretically, we prove that our prior preserves the characteristic marginal properties of global-local shrinkage priors. Moreover, we establish posterior consistency and contraction rates in growing dimensions, showing that the inclusion of structural relationships and proximal methods do not alter the shrinkage properties. Notably, these results demonstrate that the structural constraints in the model do not hinder posterior recovery. Simulation studies and real-data applications to interaction and generalised additive models demonstrate that our prior improves parameter recovery, out-of-sample prediction, and sampling efficiency over existing constrained and unconstrained shrinkage priors, even under structural misspecification.
ER  -

Abstract

Related

Cite