Visualizations for Assessing Convergence and Mixing of MCMC

Abstract

Bayesian inference often requires approximating the posterior distribution with Markov Chain Monte Carlo (MCMC) sampling. A central problem with MCMC is how to detect whether the simulation has converged. The samples come from the true posterior distribution only after convergence. A common solution is to start several simulations from different starting points, and measure overlap of the different chains. We point out that Linear Discriminant Analysis (LDA) minimizes the overlap measured by the usual multivariate overlap measure. Hence, LDA is a justified method for visualizing convergence. However, LDA makes restrictive assumptions about the distributions of the chains and their relationships. These restrictions can be relaxed by a recently introduced extension.

Cite

Text

Venna et al. "Visualizations for Assessing Convergence and Mixing of MCMC." European Conference on Machine Learning, 2003. doi:10.1007/10955628_39

Markdown

[Venna et al. "Visualizations for Assessing Convergence and Mixing of MCMC." European Conference on Machine Learning, 2003.](https://mlanthology.org/ecmlpkdd/2003/venna2003ecml-visualizations/) doi:10.1007/10955628_39

BibTeX

@inproceedings{venna2003ecml-visualizations,
  title     = {{Visualizations for Assessing Convergence and Mixing of MCMC}},
  author    = {Venna, Jarkko and Kaski, Samuel and Peltonen, Jaakko},
  booktitle = {European Conference on Machine Learning},
  year      = {2003},
  pages     = {432-443},
  doi       = {10.1007/10955628_39},
  url       = {https://mlanthology.org/ecmlpkdd/2003/venna2003ecml-visualizations/}
}