On choosing mixture components via non-local priors
From MaRDI portal
(Redirected from Publication:5213449)
Abstract: Choosing the number of mixture components remains an elusive challenge. Model selection criteria can be either overly liberal or conservative and return poorly-separated components of limited practical use. We formalize non-local priors (NLPs) for mixtures and show how they lead to well-separated components with non-negligible weight, interpretable as distinct subpopulations. We also propose an estimator for posterior model probabilities under local and non-local priors, showing that Bayes factors are ratios of posterior to prior empty-cluster probabilities. The estimator is widely applicable and helps set thresholds to drop unoccupied components in overfitted mixtures. We suggest default prior parameters based on multi-modality for Normal/T mixtures and minimal informativeness for categorical outcomes. We characterise theoretically the NLP-induced sparsity, derive tractable expressions and algorithms. We fully develop Normal, Binomial and product Binomial mixtures but the theory, computation and principles hold more generally. We observed a serious lack of sensitivity of the Bayesian information criterion (BIC), insufficient parsimony of the AIC and a local prior, and a mixed behavior of the singular BIC. We also considered overfitted mixtures, their performance was competitive but depended on tuning parameters. Under our default prior elicitation NLPs offered a good compromise between sparsity and power to detect meaningfully-separated components.
Recommendations
- scientific article; zbMATH DE number 1932865
- Mixture models with a prior on the number of components
- Bayesian Model Selection in Finite Mixtures by Marginal Density Decompositions
- A default Bayesian test for the number of components in a mixture
- On a loss-based prior for the number of components in mixture models
Cited in
(21)- Implicit models of Gaussian mixture densities and locally optimum detectors
- On posterior contraction of parameters and interpretability in Bayesian mixture modeling
- An explanatory rationale for priors sharpened into Occam's razors
- Global-local mixtures: a unifying framework
- On a loss-based prior for the number of components in mixture models
- Determinantal point process mixtures via spectral density approach
- On a class of repulsive mixture models
- Component elimination strategies to fit mixtures of multiple scale distributions
- scientific article; zbMATH DE number 1932865 (Why is no real title available?)
- Density regression using repulsive distributions
- MCMC Computations for Bayesian Mixture Models Using Repulsive Point Processes
- Heterogeneous large datasets integration using Bayesian factor regression
- Cohesion and Repulsion in Bayesian Distance Clustering
- Bayesian sparse vector autoregressive switching models with application to human gesture phase segmentation
- Bayesian clustering of high-dimensional data via latent repulsive mixtures
- Repulsion, chaos, and equilibrium in mixture models
- Distributed online expectation-maximization algorithm for Poisson mixture model
- Bayesian Wasserstein repulsive Gaussian mixture models
- On the stability of general Bayesian inference
- Posterior ramifications of prior dependence structures
- Extending local mixture models
This page was built for publication: On choosing mixture components via non-local priors
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5213449)
