Adaptation of the tuning parameter in general Bayesian inference with robust divergence
From MaRDI portal
Abstract: We introduce a methodology for robust Bayesian estimation with robust divergence (e.g., density power divergence or {gamma}-divergence), indexed by a single tuning parameter. It is well known that the posterior density induced by robust divergence gives highly robust estimators against outliers if the tuning parameter is appropriately and carefully chosen. In a Bayesian framework, one way to find the optimal tuning parameter would be using evidence (marginal likelihood). However, we numerically illustrate that evidence induced by the density power divergence does not work to select the optimal tuning parameter since robust divergence is not regarded as a statistical model. To overcome the problems, we treat the exponential of robust divergence as an unnormalized statistical model, and we estimate the tuning parameter via minimizing the Hyvarinen score. We also provide adaptive computational methods based on sequential Monte Carlo (SMC) samplers, which enables us to obtain the optimal tuning parameter and samples from posterior distributions simultaneously. The empirical performance of the proposed method through simulations and an application to real data are also provided.
Recommendations
- Robust Bayes estimation using the density power divergence
- On the `optimal' density power divergence tuning parameter
- Robust and efficient estimation by minimising a density power divergence
- Robust estimation for non-homogeneous data and the selection of the optimal tuning parameter: the density power divergence approach
- On the choice of the optimal tuning parameter in robust one-shot device testing analysis
Cites work
- A General Framework for Updating Belief Distributions
- A generalized divergence for statistical inference
- A sequential particle filter method for static models
- An introduction to sequential Monte Carlo
- An Invitation to Sequential Monte Carlo Samplers
- Bayesian model selection based on proper scoring rules
- Choosing a robustness tuning parameter
- Do robust estimators work with real data?
- Efficient strategy for the Markov chain Monte Carlo in high-dimension with heavy-tailed target probability distribution
- Estimation of non-normalized statistical models by score matching
- Families of alpha-, beta- and gamma-divergences: flexible and robust measures of similarities
- General Bayesian Loss Function Selection and the use of Improper Models
- scientific article; zbMATH DE number 194744 (Why is no real title available?)
- Laws of Large Numbers for Identically Distributed Banach-Space Valued Random Variables
- Monte Carlo maximum likelihood estimation for discretely observed diffusion processes
- On the `optimal' density power divergence tuning parameter
- On the convergence of adaptive sequential Monte Carlo methods
- On the marginal likelihood and cross-validation
- Robust and efficient estimation by minimising a density power divergence
- Robust Bayes estimation using the density power divergence
- Robust Bayesian inference via γ-divergence
- Robust parameter estimation with a small bias against heavy contamination
- Sequential Monte Carlo Samplers
- Unbiased Markov Chain Monte Carlo Methods with Couplings
Cited in
(17)- On the choice of the optimal tuning parameter in robust one-shot device testing analysis
- A reduced-rank approach to predicting multiple binary responses through machine learning
- Discussion of: ``Identifiability of latent-variable and structural-equation models: from linear to nonlinear
- Minimizing robust density power-based divergences for general parametric density models
- High-dimensional sparse classification using exponential weighting with empirical hinge loss
- Adaptive posterior concentration rates for sparse high-dimensional linear regression with random design and unknown error variance
- Sampling from density power divergence-based generalized posterior distribution via stochastic optimization
- Inequality restricted minimum density power divergence estimation in panel count data
- A robust probabilistic framework for principal component regression: optimizing parameter identification and outlier detection via approximate Bayesian computation
- On properties of fractional posterior in generalized reduced-rank regression
- Robust Bayesian graphical modeling using -divergence
- Concentration of a sparse Bayesian model with horseshoe prior in estimating high-dimensional precision matrix
- A sparse PAC-Bayesian approach for high-dimensional quantile prediction
- Concentration properties of fractional posterior in 1-bit matrix completion
- Statistical inference with non-normalized models: score matching and noise contrastive estimation
- On the stability of general Bayesian inference
- Fast robust additive models using gamma-divergence
This page was built for publication: Adaptation of the tuning parameter in general Bayesian inference with robust divergence
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6171768)
