Exact dimensionality selection for Bayesian PCA

DOI10.1111/SJOS.12424MaRDI QIDQ5108978zbMATH OpenOpenAlexFDO

Authors Pierre Latouche, Pierre-Alexandre Mattei, Charles Bouveyron

Publication date 7 May 2020

Published in Scandinavian Journal of Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1703.02834

zbMATH Keywords

dimension reductionmarginal likelihoodsingular value decompositionBayesian model selectionprincipal component analysis (PCA)

Mathematics Subject Classification ID

Bayesian inference (62F15) Factor analysis and principal components; correspondence analysis (62H25) Statistical ranking and selection procedures (62F07)

Abstract: We present a Bayesian model selection approach to estimate the intrinsic dimensionality of a high-dimensional dataset. To this end, we introduce a novel formulation of the probabilisitic principal component analysis model based on a normal-gamma prior distribution. In this context, we exhibit a closed-form expression of the marginal likelihood which allows to infer an optimal number of components. We also propose a heuristic based on the expected shape of the marginal likelihood curve in order to choose the hyperparameters. In non-asymptotic frameworks, we show on simulated data that this exact dimensionality selection approach is competitive with both Bayesian and frequentist state-of-the-art methods.

Recommendations

• Bayesian estimation of the number of principal components
• Bayesian principal component analysis with mixture priors
• Automatic PCA dimension selection for high dimensional data and small sample sizes
• scientific article; zbMATH DE number 6387599
• scientific article; zbMATH DE number 2033156

Cited in

(10)

• Condition for perfect dimensionality recovery by variational Bayesian PCA
• Bayesian estimation of the number of principal components
• Automatic PCA dimension selection for high dimensional data and small sample sizes
• scientific article; zbMATH DE number 2033156 (Why is no real title available?)
• A penalization method to estimate the intrinsic dimensionality of data
• Exploring dimension learning via a penalized probabilistic principal component analysis
• Parsimonious Gaussian mixture models with piecewise-constant eigenvalue profiles
• Scaling optimal transport to high-dimensional Gaussian distributions with application to domain adaptation
• A Bayesian vector multidimensional scaling procedure incorporating dimension reparameterization with variable selection
• Bayesian variable selection for globally sparse probabilistic PCA

Describes a project that uses

Uses Software

• GSPPCA
• Bessel

This page was built for publication: Exact dimensionality selection for Bayesian PCA

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5108978)