## Estimating Bayes factors via thermodynamic integration and population MCMC

- Ben Calderhead, Mark A. Girolami
- Computational Statistics & Data Analysis
- 2009

As mentioned in the paper Markov Chain Monte Carlo (MCMC) is a technique used to obtain samples from probability distributions known only up to a normalising factor. Given that p(~ θ) is a nonnegative integrable function, the Metropolis Hastings algorithm will provide a sequence of samples (also known as a chain) whose equilibrium distribution is… (More)