plot_predictions_vs_covariate: Plot estimated CATEs versus a covariate

In ensley-nexant/cfeval: Causal Forest Evaluation and Visualization

Description

Plot estimated CATEs versus a covariate

Usage

plot_predictions_vs_covariate(dat, res, covariate, level = NULL)

Arguments

dat	The original training data
res	The output from tidy_cf
covariate	The covariate to plot on the horizontal axis
level	The desired confidence level. If NULL, no confidence levels are plotted.

Value

A ggplot2 plot object

Examples

## Not run: 
 require(grf)

 n <- 2000; p <- 10

 X <- matrix(rnorm(n * p), n, p)
 W <- rbinom(n, 1, 0.4 + 0.2 * (X[, 1] > 0))
 Y <- pmax(X[, 1], 0) * W + X[, 2] + pmin(X[, 3], 0) + rnorm(n)
 cf <- causal_forest(X, Y, W)
 results <- tidy_cf(cf)
 dat <- dplyr::bind_cols(as.data.frame(X), results)
 dat$a <- sample(letters[1:3], size = n, replace = T)

 plot_predictions_vs_covariate(dat, 'V1', level = 0.9)
 plot_predictions_vs_covariate(dat, 'a', level = 0.9)

## End(Not run)

ensley-nexant/cfeval documentation built on May 20, 2020, 12:34 a.m.