Markov-Chain Monte Carlo (MCMC)

This subpackage provides functionality to sample posterior distributions of inverse problems in atmospheric soundings using ARTS as forward model.

The main functionality is implemented by the MCMC class which implements the Metropolis algorithm to sample from the posterior distribution.

In addition to that this subpackage provides a RandomWalk class that simplifies the setup of random walk jump functions as well as diagnostic function to assess mixing and convergence of the simulations.

MCMC(vars, y, ly[, stats])

The MCMC class represents an ongoing MCMC simulation.