Exact Dimensionality Selection for Bayesian PCA

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.
Liste complète des métadonnées

Contributor : Pierre-Alexandre Mattei <>
Submitted on : Monday, March 6, 2017 - 5:59:38 PM
Last modification on : Thursday, April 11, 2019 - 4:02:09 PM
Document(s) archivé(s) le : Wednesday, June 7, 2017 - 4:00:23 PM


Files produced by the author(s)


  • HAL Id : hal-01484099, version 1
  • ARXIV : 1703.02834


Charles Bouveyron, Pierre Latouche, Pierre-Alexandre Mattei. Exact Dimensionality Selection for Bayesian PCA. 2017. ⟨hal-01484099⟩



Record views


Files downloads