# Miscellaneous

I sometimes play around with different idea/samplers/algorithms. I hope to keep adding articles to this page.

- Fast sampling with Gaussian scale-mixture priors in high-dimensional regression

A handy algorithm that lets one sample from a multivariate normal with a specific mean and variance structure at a faster rate than the usual sampling procedure. This is when the parameters in the multivariate normal require a matrix inverse, and can be used in Bayesian lasso style problems.

R Markdown Implenetation - Regenerative Simulation: Bayesian Probit Regression

Here I implement regenerative simulation for the Bayesian probit regression model, following the steps in the paper. The purpose of this was to implement regenerative simulation for a problem since I have not seen a detailed explanation of implementation.

R Markdown Implenetation - MALA versus Random Walk Metropolis

As an exercise I implement the Metropolis-adjusted Langevin algorithm (MALA) for two simple target distributions and compare their performance with the random walk Metropolis.

R Markdown Implenetation