Ancillarity-sufficiency interweaving strategy (ASIS) for boosting MCMC estimation of stochastic volatility models

• Published in: Computational statistics & data analysis Vol. 76; pp. 408 - 423
• Main Authors: Kastner, Gregor, Frühwirth-Schnatter, Sylvia
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.08.2014
• Subjects: Auxiliary mixture sampling; Bayesian theory; equations; Exchange rate data; Inference; Markov chain Monte Carlo; Massively parallel computing; Mathematical models; Non-centering; Parametrization; Sampling; State space model; Statistics; Stochasticity; Strategy; Volatility; State space model; Exchange rate data; Auxiliary mixture sampling; Markov chain Monte Carlo; Massively parallel computing; Non-centering

Bayesian inference for stochastic volatility models using MCMC methods highly depends on actual parameter values in terms of sampling efficiency. While draws from the posterior utilizing the standard centered parameterization break down when the volatility of volatility parameter in the latent state...