In my last post I provided a small list of some R packages for random forest. Today I will provide a more complete list of random forest R packages. In the first table I list the R packages which contains the possibility to perform the standard random forest like described in the original Breiman paper.

xgboost is a package that is used for boosted tree algorithms and is the current state of the art for machine learning challenges for example on the platform Kaggle due to its flexibility and very good performance. xgboost also contains the possibility to grow a random forest, as can be seen in the last section of this tutorial page.

A graphical representation of this table:

Three algorithms that are especially constructed for big datasets are the following three (ranger and Rborist could also have been listed here):

bigrf is currently not on CRAN anymore as check problems were not corrected despite reminders, that is the reason for the low download statistic.

An interesting blog post from Szilard about benchmarking different packages (also on other software like Python and only one dataset) on big datasets and classification tasks can be found here. It is mainly interesting for seeing the differences in speed performance. Benchmarking performance on only one dataset is clearly not enough. The github page of the whole study can be seen here

Many more packages exist which are slightly different from the original random forest algorithm. These are listed in the following table:

rpart is just for creating a single tree. party contains cforest which are random forests based on conditional inference trees.

R-code for obtaining the tables and the graphic:

library(cranlogs)
library(data.table)
library(knitr)

downloads = cran_downloads(packages = c("randomForest", "xgboost", "randomForestSRC", "ranger", "Rborist"), when = c("last-month"))
downloads = data.table(downloads)
downloads = downloads[,sum(count), by = "package"]
colnames(downloads) = c("**package**", "**RStudio downloads in the last month**")
kable(downloads, format = "markdown")
barplot(downloads[,2], names.arg = downloads$package, col = "blue")

downloads = cran_downloads(packages = c("h2o","ParallelForest", "bigrf"), when = c("last-month"))
downloads = data.table(downloads)
downloads = downloads[,sum(count), by = "package"]
colnames(downloads) = c("**package**", "**RStudio downloads in the last month**")
kable(downloads, format = "markdown")

downloads = cran_downloads(packages = c("rpart", "RRF", "randomForestSRCSyn", "obliqueRF", "rotationForest", 
                                        "rFerns", "randomUniformForest", "wsrf", "roughrf", "trimTrees", "extraTrees", "party" ), when = c("last-month"))
downloads = data.table(downloads)
downloads = downloads[,sum(count), by = "package"]
downloads = downloads[order(downloads$V1, decreasing = T),]
colnames(downloads) = c("**package**", "**RStudio downloads in the last month**")
kable(downloads, format = "markdown")