Is there an R random forest implementation that works well with very sparse data? I have thousands or millions of boolean input variables, but only hundreds or so will be TRUE for any given example.
I'm relatively new to R and noticed that there is a 'Matrix' package for dealing with sparse data, but the standard 'randomForest' package doesn't seem to recognize this data type. If it matters, the input data is going to be produced outside of R and imported.
Any advice? I can also look into using Weka, Mahout or other packages.