bangecon/bsPDP: Bootstrapped PDPs with Confidence Intervals

This is a collection of functions for inference machine learning and other methods using partial dependence plots. bsPDP() to bootstrap a predictive model for the purpose of computing confidence intervals. bsPDP.lm() implements bsPDP for simple least-squares linear models bsPDP.glm() implements bspdp for generalized linear models bsPDP.randomForest() implements bsPDP with the randomForest package bsPDP.caret() implements bsPDP with the caret package (and corresponding methods) bsPDP.BayesTree() and bsPDP.dbarts implements bsPDP for BART (dbarts is faster). bsPDP.iptw_est() for the inverse probability of treatment weights estimator. bsPDP.hi_est() for the Hirano and Imbens (2004) estimator. plot_bsPDP() to plot the bootstrapped PDPs.

install.packages("remotes")
remotes::install_github("bangecon/bsPDP")