Final published version
Other version
Research output: Contribution to Journal/Magazine › Journal article › peer-review
Research output: Contribution to Journal/Magazine › Journal article › peer-review
}
TY - JOUR
T1 - Two-level stochastic search variable selection in GLMs with missing predictors
AU - Mitra, Robin
AU - Dunson, David
PY - 2010/1/1
Y1 - 2010/1/1
N2 - Stochastic search variable selection (SSVS) algorithms provide an appealing and widely used approach for searching for good subsets of predictors while simultaneously estimating posterior model probabilities and model-averaged predictive distributions. This article proposes a two-level generalization of SSVS to account for missing predictors while accommodating uncertainty in the relationships between these predictors. Bayesian approaches for allowing predictors that are missing at random require a model on the joint distribution of the predictors. We show that predictive performance can be improved by allowing uncertainty in the specification of predictor relationships in this model. The methods are illustrated through simulation studies and analysis of an epidemiologic data set.
AB - Stochastic search variable selection (SSVS) algorithms provide an appealing and widely used approach for searching for good subsets of predictors while simultaneously estimating posterior model probabilities and model-averaged predictive distributions. This article proposes a two-level generalization of SSVS to account for missing predictors while accommodating uncertainty in the relationships between these predictors. Bayesian approaches for allowing predictors that are missing at random require a model on the joint distribution of the predictors. We show that predictive performance can be improved by allowing uncertainty in the specification of predictor relationships in this model. The methods are illustrated through simulation studies and analysis of an epidemiologic data set.
KW - missing at random
KW - model averaging
KW - multiple imputation
KW - stochastic search
KW - subset
KW - variable selection
U2 - 10.2202/1557-4679.1173
DO - 10.2202/1557-4679.1173
M3 - Journal article
VL - 6
JO - International Journal of Biostatistics
JF - International Journal of Biostatistics
IS - 1
ER -