One of my friends told me that I can use Hamiltonian Monte Carlo (HMC) to estimate parameters of Gaussian process regression models, but I am new to Stan and don’t understand how it works. From what I read on this website, HMC is a tool for sampling. How can this sampling tool be used to estimate parameters for regression model? Thanks!
HMC is the default sampler in Stan, so you need to learn how to code your model in the Stan language, then behind the scenes Stan will use HMC to sample the posterior. For GPs specifically, see the Chapter in the manual: https://mc-stan.org/docs/2_20/stan-users-guide/gaussian-processes-chapter.html
1 Like