# Category Archives: Uncategorized

## Why and how pseudo-marginal MCMC work for exact Bayesian inference

I will describe a breakthrough strategy for “exact-approximate” Bayesian inference. The apparent contradiction in the terminology is due to the surprising result in Beaumont (2003) and Andrieu and Roberts (2009) where it is shown that plugging a non-negative and unbiased … Continue reading

## Tips for coding a Metropolis-Hastings sampler

I will suggest several tips, and discuss common beginner’s mistakes occuring when coding from scratch a Metropolis-Hastings algorithm. I am making this list from the top of my mind, so feel free to propose suggestions by commenting to this post. … Continue reading

## Sequential Monte Carlo and the bootstrap filter

In the previous post I have outlined three possibilities for approximating the likelihood function of a state space model (SSM). That post is a required read to follow the topics treated here. I concluded with sequential importance sampling (SIS), which … Continue reading

## Monte Carlo sampling for likelihood approximation in state space models

In a previous post I have set the problem of estimating parameters for a state-space model (SSM). That post is a required read, also because I set some notation. My final goal is to show how to construct exact Bayesian … Continue reading

## State space models and intractable likelihoods

In this first series of posts, I introduce important tools to construct inference methods for the estimation of parameters in stochastic models. Stochastic models are characterized by randomness in their mathematical nature, and since at first I focus on models … Continue reading

## Welcome!

Welcome to my first blog post! You can read about me in the About section. I will write about statistical inference methods and algorithms, typically (though not exclusively) for models that have some dynamic component. Posts will reflect personal research … Continue reading