Intuitive joint priors for high-dimensional Bayesian linear models

dc.contributor.advisorBürkner, Paul-Christian
dc.contributor.authorAguilar Romero, Javier Enrique
dc.contributor.refereeIckstadt, Katja
dc.contributor.refereeVehtari, Aki
dc.date.accepted2026-03-25
dc.date.accessioned2026-11-17T08:19:46Z
dc.date.issued2026
dc.description.abstractIn high-dimensional Bayesian regression, prior distributions can decisively affect posterior inference. These models often involve multilevel terms, grouped predictors, interactions, or dependent covariates, yet standard shrinkage priors are typically formulated for exchangeable coefficients and do not directly encode this structure. Moreover, coefficient-level hyperparameters can obscure how prior choices affect predictive behavior, making it difficult to incorporate domain knowledge. To address these problems, we develop a predictive-first framework in which we specify priors on the coefficient of determination R2 and use variance decompositions to translate these specifications into joint shrinkage priors on model parameters. Our construction induces regularization across and within structured components of the linear predictor. We extend this framework to multilevel models, grouped predictors, and dependent covariates. We show that the resulting prior families have desirable shrinkage properties and introduce interpretable quantities that help users understand how prior choices affect regularization. We evaluate the proposed priors in simulation studies and real-data applications, examining when and to what extent incorporating structure improves parameter recovery and out-of-sample prediction.en
dc.identifier.urihttp://hdl.handle.net/2003/45269
dc.identifier.urihttp://dx.doi.org/10.17877/DE290R-27037
dc.language.isoen
dc.subjectBayesian inferenceen
dc.subjectBayesian workflowen
dc.subjectprior specificationen
dc.subjectshrinkage priorsen
dc.subjecthigh-dimensional regressionen
dc.subject.ddc310
dc.subject.rswkBayes-Verfahrende
dc.subject.rswkRegressionsmodellde
dc.subject.rswkPrädiktorde
dc.subject.rswkA-priori-Verteilungde
dc.titleIntuitive joint priors for high-dimensional Bayesian linear modelsen
dc.typeText
dc.type.publicationtypePhDThesis
dcterms.accessRightsopen access
eldorado.dnb.deposittrue
eldorado.secondarypublicationfalse

Dateien

Originalbündel

Gerade angezeigt 1 - 1 von 1
Lade...
Vorschaubild
Name:
Dissertation_Aguilar.pdf
Größe:
15.61 MB
Format:
Adobe Portable Document Format
Beschreibung:
DNB

Lizenzbündel

Gerade angezeigt 1 - 1 von 1
Lade...
Vorschaubild
Name:
license.txt
Größe:
4.82 KB
Format:
Item-specific license agreed upon to submission
Beschreibung: