RFCLUST

RFCLUST performs Random Forests of Divisive Monothetic (divclust) Trees for Unsupervised Clustering.

Installation

You can install the development version of RFCLUST from GitHub.

RFCLUSTdepends on a custopmized implementation of the divclust package, that must first be installed with the following:

# install.packages("remotes")
remotes::install_github("Alar0g/divclust")

Then, RFCLUST can be installed with:

remotes::install_github("Alar0g/RFCLUST")

Example

library(RFCLUST)
#> Loading required package: parallel
#> Registered S3 method overwritten by 'GGally':
#>   method from   
#>   +.gg   ggplot2
library(palmerpenguins)
mypeng <- as.data.frame(penguins)
mypeng$year <- factor(as.character(mypeng$year),
                         levels=c("2007", "2008", "2009"),
                         ordered=TRUE)

forest_clust <- rfclust(na.omit(mypeng[mypeng$sex=="male", -c(1, 7)]), ntrees = 50, ncores = 1)
#> We advise you do use the `summary()` on this object to agregate the result of this forest, before plotting the summary itself.

resume <- summary(forest_clust)
plot(resume)