B1222
Title: Two-component Gibbs samplers: Convergence rate and asymptotic variance
Authors: Qian Qin - University of Minnesota (United States) [presenting]
Galin Jones - University of Minnesota (United States)
Abstract: Deterministic-scan and random-scan Gibbs samplers with two components are compared. In terms of convergence rate, the deterministic-scan version is superior. On the other hand, in terms of asymptotic variance, the random-scan version is more robust. The comparison takes into account the computation cost of the MCMC algorithms.
