Björn-Hergen Laabs
timbR is a collection of methods for the interpretation of random forests trained by ranger.
In version 1.0 most representaive trees can be selected based on four different tree based distance measures.
To install the timbR R package, run
install.packages("devtools")
library(devtools)
devtools::install_github("imbs-hl/timbR")
If you find any bugs, or if you experience any crashes, please report to us.
Please cite our paper if you use ranger.
- Laabs, B.-H., Westenberger, A., König, I. R. (2020) Identification of representative trees in random forests based on a new tree-based distance measure. Unpublished
- Bannerjee, M., Ding, Y., Noone, A.-M. (2012) Identifying representative trees from ensembles. Stat in Med 31:1601-16.