Efficient Markov chain Monte Carlo with incomplete multinomial data

Multinomial Data By KWANG WOO AHN and KUNG-SIK CHAN Department of Statistics and Actuarial Science, The University of Iowa, Iowa City, IA USA [email protected] [email protected] Summary We propose a new, block Gibbs sampling scheme for incomplete multinomial data. The new approach facilitates maximal blocking, thereby reducing serial dependency and speeding up the convergence of the Gibbs sampler. We compare the new method with the standard, non-block Gibbs sampler via a numerical example. Some key words : Blocking; Gibbs Sampler; Dirichlet distribution; Epidemiology.