Sufficient Conditions for Ergodicity and Convergence of MH, SA, and EM Algorithms

Autor: D. S. B. Martins Neto, C. C. Y. Dorea, A. G. C. Pereira
Rok vydání: 2004
Předmět:
Zdroj: Stochastic Models. 20:193-204
ISSN: 1532-4214
1532-6349
DOI: 10.1081/stm-120034128
Popis: A simple sufficient condition for ergodicity of Markov chains that possess transition kernel of the type P(x, A) = ∫ A p(x, y)ν(dy) + r(x)1 A (x) is presented. Applications includes the convergence of several MCMC algorithms: Metropolis-Hastings, Simulated Annealing and Expectation Maximization (EM).
Databáze: OpenAIRE
Nepřihlášeným uživatelům se plný text nezobrazuje