fuzzyforest is an extension of random forests designed to yield less biased variable importance rankings when features are correlated with one another. The algorithm requires that features be partitioned into separate groups or modules such that the correlation within groups is large and the correlation between groups is small. fuzzyforest allows for easy integration with the package WGCNA.

• the latest released version can be downloaded from CRAN: install.packages("fuzzyforest")
• the latest development version can be downloaded from github: install_github("OHDSI/FuzzyForest")

To enable use of the full functionality of fuzzyforest, the package WGCNA must be installed. However, WGCNA requires the installation of a few packages from bioConductor. To install WGCNA, type the following lines into the console:

setRepositories(ind=1:2)  
install.packages("WGCNA")
source("http://bioconductor.org/biocLite.R")
biocLite("AnnotationDbi", type="source")
biocLite("GO.db")

If further issues with the installation of WGCNA arise see the WGCNA website: http://labs.genetics.ucla.edu/horvath/CoexpressionNetwork/Rpackages/WGCNA/index.html#manualInstall

This work is partially supported through NSF grant IIS 1251151.