Implementation of a statistical approach for estimating the joint health effects of multiple concurrent exposures.
The R package
bkmr implements Bayesian kernel machine regression, a statistical approach for estimating the joint health effects of multiple concurrent exposures. Additional information on the statistical methodology and on the computational details are provided in Bobb et al. 2015.
You can install the latest releasted version of
bkmr from CRAN with:
Or the latest development version from github with:
For a general overview and guided examples, go to https://jenfb.github.io/bkmr/overview.html.
Added ability to have binomial outcome
family by implementing probit regression within
Removed computation of the subject-specific effects
kmbayes(), as this is not always desired, and greatly slows down model fitting
This could still be done by setting the option
est.h = TRUE in the
posterior samples of
h[i] can now be obtained via the post-processing
SamplePred function; alternatively, posterior summaries (mean, variance) can be obtained via the post-processing
Added ability to use exact estimates of the posterior mean and variance by specifying the argument
method = 'exact' within the post-processing functions (e.g.,
both_pairs = TRUE(#4)