|
Abstract:
|
Bayesian Model Averaging (BMA) has emerged as a useful model-building methodology over the past several decades. It may be especially useful to researchers who are faced with a large number of potential predictor variables of unknown utility in answering the study questions at hand. One benefit of BMA is that it does not require the analyst to apriori choose a subset of predictor variables, but rather allows for each variable to be weighted according to its posterior probability of being included in the model. We explore the features of BMA on two real public health datasets are aimed at studying new therapies for treating patients with latent Tuberculosis infection and HIV, respectively.
|
