Skip to contents

Basic information

Title: BKMR

Package Author: Jennifer F. Bobb

Website Author: Sunan Gao

Description: 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. More recent extensions, details on the software, and worked-through examples are provided in Bobb et al. 2018.

Original R package came from Link

Deployed Website Link ### Customized things in pkgdown website

  theme: breeze-light  # using the "breeze-light" theme
  bslib:
    bg: '#F5F5F5' # Specifies the background color for the website.
    fg: '#B8BCC2' # Specifies the foreground (text) color for the website.
    primary: '#306cc9' # Specifies the primary color used for links and other elements.
  navbar:
    bg: primary # Sets the background color of the navigation bar
    structure:
      left: search # Places a search bar on the left side of the navigation bar.
      right:
      - reference
      - articles
  footer:
    structure:
      left: developed_by # Specifies content on the left side of the footer
      right: built_with # Specifies content on the right side of the footer

Exported function list 

ComputePostmeanHnew():  #Compute the posterior mean and variance of h at a
new predictor values

ExtractEsts(): #Extract summary statistics

ExtractPIPs(): #Extract posterior inclusion probabilities (PIPs) from
BKMR model fit

ExtractSamps(): #Extract samples

InvestigatePrior(): #Investigate prior

OverallRiskSummaries(): #Calculate overall risk summaries

PlotPriorFits(): #Plot of exposure-response function from univariate KMR
fit

PredictorResponseBivar(): #Predict the exposure-response function at a
new grid of points

PredictorResponseBivarLevels(): #Plot cross-sections of the bivariate predictor-response function

PredictorResponseBivarPair(): #Plot bivariate predictor-response function on a new grid of points

PredictorResponseUnivar(): #Plot univariate predictor-response function on a new grid of points

SamplePred(): #Obtain posterior samples of predictions at new points
SimData(): #Simulate dataset

SingVarIntSummaries(): #Single Variable Interaction Summaries

SingVarRiskSummaries(): #Single Variable Risk Summaries

TracePlot(): #Trace plot

kmbayes(): #Fit Bayesian kernel machine regression

print(<bkmrfit>): #Print basic summary of BKMR model fit

summary(<bkmrfit>): #Summarizing BKMR model fits

A basic example with one of the functions (fitkm). 

library(bkmr)
library(ggplot2)

set.seed(111)
dat <- SimData(n = 50, M = 4)
y <- dat$y
Z <- dat$Z
X <- dat$X
set.seed(111)
fitkm <- kmbayes(y = y, Z = Z, X = X, iter = 10000, verbose = FALSE, varsel = TRUE)
TracePlot(fit = fitkm, par = "beta")
ggplot(pred.resp.univar, aes(z, est, ymin = est - 1.96*se, ymax = est + 1.96*se)) + 
    geom_smooth(stat = "identity") + 
    facet_wrap(~ variable) +
  ylab("h(z)")

Install instructions

You can install the latest released version of bkmr from CRAN with:

Or the latest development version from github with:

install.packages("devtools")
devtools::install_github("jenfb/bkmr")

For a general overview and guided examples, go to https://jenfb.github.io/bkmr/overview.html.

For examples from the software paper, please see