Posterior Summaries for Nonparametric Bayesian Mixture Models Using Optimal Transport

A methodological paper and associated data propose a novel approach for summarizing posterior inference in nonparametric Bayesian mixture models. The work, authored by Khai Nguyen and available on figshare, was last updated in April 2026. It introduces two variants of sliced Wasserstein distance for Gaussian mixtures to estimate mixing measures and validate partitions.

Use Cases

• Comparing posterior summaries for clustering based on the described sliced optimal transport metrics.
• Validating density estimates from mixture models using the proposed decision-theoretic framework.
• Implementing the mixed sliced Wasserstein or sliced mixture Wasserstein distances for Gaussian mixture models.

Strengths

• The methodology is described as model-agnostic, remaining valid under complex dependence structures.
• The dataset is licensed under CC-BY-4.0, allowing for open sharing and adaptation.
• The associated files have a defined size of 16.5 MB.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The description is methodological; the nature and scale of any accompanying numerical data are unspecified.

Provenance

Source

figshare

Freshness

Last updated 2026-04-21 19:22:31; freshness should be verified.