Metropolis–Hastings algorithm
Algoritmo de Metropolis-Hastings
- Date
- 1953 (Metropolis) / 1970 (generalización de Hastings) · confidence medium
- Place
- Los Alamos, New Mexico, United States
- Credited to
- W. K. Hastings, Stanisław Ulam, Edward Teller, Enrico Fermi, Nicholas Metropolis y Marshall Rosenbluth
- Field
- Knowledge & symbols
What it is
A Markov chain Monte Carlo (MCMC) method that draws samples from a probability distribution that is hard to sample directly, accepting or rejecting each proposed sample based on its relative probability.
The original Metropolis algorithm (1953, Los Alamos) was generalized by W. K. Hastings in 1970, in Toronto.
What it led to
- efficient sampling in statistical physics and biostatistics
- the foundation of modern computational Bayesian inference
- extension to complex multidimensional distributions
- widespread use in probabilistic machine learning