ICEbox

ICEbox provides tools for constructing and visualizing Individual Conditional Expectation (ICE) plots. ICE plots show the fitted relationship between a feature and the predicted response for individual observations, revealing heterogeneity and interactions that an averaged partial dependence plot can hide.

For the method and worked examples, see Goldstein, Kapelner, Bleich, and Pitkin (2015), “Peeking Inside the Black Box: Visualizing Statistical Learning With Plots of Individual Conditional Expectation”, or the freely available arXiv version.

Installation

Install the CRAN release with:

install.packages("ICEbox")

Or install the latest build from Adam Kapelner’s R-universe:

install.packages(
  "ICEbox",
  repos = c(
    kapelner = "https://kapelner.r-universe.dev",
    CRAN = "https://cloud.r-project.org"
  )
)

Basic use

library(ICEbox)

fit <- lm(mpg ~ wt + hp + disp, data = mtcars)
X <- mtcars[c("wt", "hp", "disp")]

ice_fit <- ice(fit, X = X, y = mtcars$mpg, predictor = "wt", verbose = FALSE)
plot(ice_fit)