CODEWITHSUNDEEP
rmcmc
Adrian Santos
Adrian Santos

Reputation: 113

MCMCglmm Bayes Factor

I am trying to fit by means of the MCMCglmm package a Bayesian model for fitting my data. However, I cannot find any reference on how to calculate the Bayes Factor for comparing two models.

Any suggestions or ideas?

Upvotes: 2

Views: 559

Answers (1)

zozi9126
zozi9126

Reputation: 11

If I understand correctly, You'd like to do model comparison - in this case, maybe DIC (deviance information criterion; in the MuMIn R-package) should be enough. In R You can use DIC(model1, model2) to get DIC scores for the models. It works similarly to other information criterion scores (such as AIC), hence the lower scores are 'better', but DIC is more appropriate for models in which estimates are obtained by Markov Chain Monte Carlo (MCMC) simulations. BIC (Bayesian information criterion; in VGAM R-package) has similar properties.

I hope this helps.

Cheers,

Zoltan

Upvotes: 1

Related Questions