| MCMC_LOGIT_KEEP {VARSELECTEXPOSURE} | R Documentation |
Obtains posterior samples from an MCMC algorithm to perform variable selection.
Description
Performs posterior sampling from an MCMC algorithm to estimate average treatment effect and posterior probability of inclusion of candidate variables.
Usage
MCMC_LOGIT_KEEP(Y, Z, PIN, MAX_COV, SdBeta, NUM_REPS)
Arguments
Y |
Binary outcome vector. |
Z |
Matrix of covariates including binary exposure variable. |
PIN |
Prior probability of inclusion of candidate variables. |
MAX_COV |
Maximum number of covariates in desired model. |
SdBeta |
Prior standard deviation for generating distrubtion of proposal coefficients. |
NUM_REPS |
Number of MCMC iterations to perform. |
Value
List containing (1) the posterior distribution of the estimated Average Treatment Effect, (2) the posterior distributions of the intercept parameter, (3) the posterior distributions of the rest of the coefficients including the exposure coefficient, and (4) the posterior distribution for the indication of whether or not the variable was included in a given iteration's model.