A New, Conditional Variable-Importance Measure for Random Forests Available in the party Package
Carolin Strobl, Torsten Hothorn, Achim Zeileis · 2009
Random forests are one of the most popular statistical learning algorithms, and a variety of methods for fitting random forests and related recursive partitioning approaches is available in R. This paper points out two impor- tant features of the random forest implementa- tion cforest available in the party package: The resulting forests are unbiased and thus prefer- able to the randomForest implementation avail- able in randomForest if predictor variables are of different types. Moreover, a conditional per- mutation importance measure has recently been added to the party package, which can help eval- uate the importance of correlated predictor vari- ables. The rationale of this new measure is illus- trated and hands-on advice is given for the usage of recursive partitioning tools in R.