Open AccessPosted Content
Bayesian Covariance Matrix Estimation using a Mixture of Decomposable Graphical Models
TLDR
It is shown empirically that the prior that assigns equal probability over graph sizes outperforms the prior over all graphs in more efficiently estimating the covariance matrix.Abstract:
A Bayesian approach is used to estimate the covariance matrix of Gaussian data. Ideas from Gaussian graphical models and model selection are used to construct a prior for the covariance matrix that is a mixture over all decomposable graphs. For this prior the probability of each graph size is specified by the user and graphs of equal size are assigned equal probability. Most previous approaches assume that all graphs are equally probable. We show empirically that the prior that assigns equal probability over graph sizes outperforms the prior that assigns equal probability over all graphs, both in identifying the correct decomposable graph and in more efficiently estimating the covariance matrix.read more
Citations
More filters
Book
Machine Learning : A Probabilistic Perspective
TL;DR: This textbook offers a comprehensive and self-contained introduction to the field of machine learning, based on a unified, probabilistic approach, and is suitable for upper-level undergraduates with an introductory-level college math background and beginning graduate students.
Journal ArticleDOI
Bayesian inference for general gaussian graphical models with application to multivariate lattice data
TL;DR: This work introduces efficient Markov chain Monte Carlo methods for inference and model determination in multivariate and matrix-variate Gaussian graphical models and extends their sampling algorithms to a novel class of conditionally autoregressive models for sparse estimation inMultivariate lattice data.
Journal ArticleDOI
Efficient Bayesian Inference for Multivariate Probit Models With Sparse Inverse Correlation Matrices
TL;DR: A Bayesian approach for inference in the multivariate probit model, taking into account the association structure between binary observations, is proposed and an efficient Markov chain Monte Carlo algorithm relying on a parameter expansion scheme to sample from the resulting posterior distribution is proposed.
Journal ArticleDOI
Sparse covariance estimation in heterogeneous samples
TL;DR: In this article, the authors explore mixtures of Gaussian graphical models and infinite hidden Markov models where the emission distributions correspond to Gaussian graph models, allowing them to divide a heterogeneous population into homogeneous groups, with each cluster having its own conditional independence structure.
Journal ArticleDOI
Robust Bayesian Graphical Modeling Using Dirichlet $t$-Distributions
Michael Finegold,Mathias Drton +1 more
TL;DR: Shahbaba et al. as discussed by the authors proposed the Dirichlet t-distribution, which combines more appealing modeling of the multivariate dependence structure with favorable computational properties, and used the Gaussian model to extract information from highly multivariate data even when most experiments contain outliers for some of their measurements.
References
More filters
Book
Aspects of multivariate statistical theory
TL;DR: In this paper, the authors present a set of standard tests on Covariance Matrices and Mean Vectors, and test independence between k Sets of Variables and Canonical Correlation Analysis.
Journal ArticleDOI
Markov Chains for Exploring Posterior Distributions
TL;DR: Several Markov chain methods are available for sampling from a posterior distribution as discussed by the authors, including Gibbs sampler and Metropolis algorithm, and several strategies for constructing hybrid algorithms, which can be used to guide the construction of more efficient algorithms.
Journal ArticleDOI
Mixtures of g Priors for Bayesian Variable Selection
TL;DR: In this paper, a mixture of g priors is proposed as an alternative to the default g prior, which resolves many of the problems with the original formulation while maintaining the computational tractability of the g prior.
Journal ArticleDOI
Markov Chain Monte Carlo in Practice: A Roundtable Discussion
TL;DR: In this paper, the authors present advice and guidance to novice users of MCMC and not-so-novice users as well, including building confidence in simulation results, methods for speeding and assessing convergence, estimating standard error, etc.
Journal Article
Modeling covariance matrices in terms of standard deviations and correlations, with application to shrinkage
TL;DR: A statistically motivated decomposition which appears to be relatively unexplored for the purpose of modeling is studied, and a straightforward computational strategy for obtaining the posterior of the covariance matrix is described.