To MCMC or not to MCMC: Evaluating non-MCMC methods for Bayesian penalized regression
Published in arXiv preprint, 2025
In this paper we compare MCMC and non-MCMC methods for high-dimensional penalized regression, using a simulation study and empirical datasets with continuous and binary outcomes. Mean-field variational inference often performs comparably to MCMC at a fraction of the runtime, but the trade-off depends on the context. The paper also includes a step-by-step tutorial in R.
Recommended citation: van Leeuwen, F. D., & van Erp, S. (2025). "To MCMC or not to MCMC: Evaluating non-MCMC methods for Bayesian penalized regression." arXiv preprint, arXiv:2510.20947.
Download Paper