Bayesian predictive model comparison via parallel sampling Export

@article{citeulike:3575867, abstract = {Methods of model comparison and checking, and associated criteria, are proposed based on parallel sampling of two or more models subsequent to convergence. These complement Bayesian predictive criteria already proposed (e.g. error sum of squares and deviance based) but are on a scale that may be compared across applications. Penalised criteria for model comparison based on the AIC are also investigated, together with AIC model weights and evidence ratios. Parallel sampling enables posterior summaries to be obtained for continuous comparison measures (e.g. likelihood and evidence ratios). A forward selection procedure for regression is suggested as one possible extension, as well as procedures for model averaging and posterior predictive checking. Comparisons with the DIC are made together with implications of parallel sampling for assessing the density of the DIC. Three worked examples illustrate the working of the procedures in practice.}, author = {Congdon, P.}, citeulike-article-id = {3575867}, citeulike-linkout-0 = {http://dx.doi.org/10.1016/j.csda.2004.03.016}, citeulike-linkout-1 = {http://linkinghub.elsevier.com/retrieve/pii/S0167947304000982}, day = {01}, doi = {10.1016/j.csda.2004.03.016}, issn = {01679473}, journal = {Computational Statistics \& Data Analysis}, keywords = {bayesian, mcmc, model-selection, prediction, sampling, statistics}, month = {April}, number = {4}, pages = {735--753}, posted-at = {2009-10-31 15:37:01}, priority = {2}, title = {Bayesian predictive model comparison via parallel sampling}, url = {http://dx.doi.org/10.1016/j.csda.2004.03.016}, volume = {48}, year = {2005} }

TY - JOUR ID - citeulike:3575867 L3 - citeulike-article-id:3575867 N2 - Methods of model comparison and checking, and associated criteria, are proposed based on parallel sampling of two or more models subsequent to convergence. These complement Bayesian predictive criteria already proposed (e.g. error sum of squares and deviance based) but are on a scale that may be compared across applications. Penalised criteria for model comparison based on the AIC are also investigated, together with AIC model weights and evidence ratios. Parallel sampling enables posterior summaries to be obtained for continuous comparison measures (e.g. likelihood and evidence ratios). A forward selection procedure for regression is suggested as one possible extension, as well as procedures for model averaging and posterior predictive checking. Comparisons with the DIC are made together with implications of parallel sampling for assessing the density of the DIC. Three worked examples illustrate the working of the procedures in practice. IS - 4 JF - Computational Statistics & Data Analysis SN - 01679473 EP - 753 TI - Bayesian predictive model comparison via parallel sampling VL - 48 SP - 735 KW - bayesian KW - mcmc KW - model-selection KW - prediction KW - sampling KW - statistics AU - Congdon, P PY - 2005/04/01/ UR - http://dx.doi.org/10.1016/j.csda.2004.03.016 ER -