This package is DEPRECATED. Please use the packages `grf` or `ranger` instead, which have built-in confidence intervals.
R
Switch branches/tags
Nothing to show
Latest commit 713d6ed Dec 12, 2017 @swager swager Update README.md

README.md

randomForestCI

πŸ”΄ This package is deprecated. Please use one of the following packages instead: πŸ”΄

  • grf, which has built-in support for resampling-based confidence intervals, or
  • ranger, which has an actively maintained version of the infinitesimal jackknife for random forests.

πŸ”΄ Both packages are available from CRAN. πŸ”΄

Confidence intervals for random forests using the infinitesimal jackknife, as developed by Efron (2014) and Wager et al. (2014).

To install this package in R, run the following commands:

install.packages("devtools")
library(devtools) 
install_github("swager/randomForestCI")

Example usage:

library(randomForestCI)

# Make some data...
n = 250
p = 100
X = matrix(rnorm(n * p), n, p)
Y = rnorm(n)
  
#  Run the method
rf = randomForest(X, Y, keep.inbag = TRUE)
ij = randomForestInfJack(rf, X, calibrate = TRUE)

plot(ij)

References

Efron, Bradley. Estimation and accuracy after model selection. Journal of the American Statistical Association, 109(507), 2014. [link]

Wager, Stefan, Trevor Hastie, and Bradley Efron. Confidence intervals for random forests: The jackknife and the infinitesimal jackknife. The Journal of Machine Learning Research, 15(1), 2014. [link]