Previous: Additional Inputs

Up: mlogit.bayes: Bayesian Multinomial Logistic

Next: Examples

Convergence

Users should verify that the Markov Chain converges to its stationary distribution. After running the zelig() function but before performing setx(), users may conduct the following convergence diagnostics tests:

• geweke.diag(z.out$coefficients): The Geweke diagnostic tests the null hypothesis that the Markov chain is in the stationary distribution and produces z-statistics for each estimated parameter.

• heidel.diag(z.out$coefficients): The Heidelberger-Welch diagnostic first tests the null hypothesis that the Markov Chain is in the stationary distribution and produces p-values for each estimated parameter. Calling heidel.diag() also produces output that indicates whether the mean of a marginal posterior distribution can be estimated with sufficient precision, assuming that the Markov Chain is in the stationary distribution.

• raftery.diag(z.out$coefficients): The Raftery diagnostic indicates how long the Markov Chain should run before considering draws from the marginal posterior distributions sufficiently representative of the stationary distribution.

If there is evidence of non-convergence, adjust the values for burnin and mcmc and rerun zelig().

Advanced users may wish to refer to help(geweke.diag), help(heidel.diag), and help(raftery.diag) for more information about these diagnostics.