Most Popular Learners in mlr

An attempt to asses the popularity of mlr learners
R
r-bloggers
Author

Jabok Richter

Published

March 30, 2017

For the development of mlr as well as for an “machine learning expert” it can be handy to know what are the most popular learners used. Not necessarily to see, what are the top notch performing methods but to see what is used “out there” in the real world. Thanks to the nice little package cranlogs from metacran you can at least get a slight estimate as I will show in the following…

First we need to install the cranlogs package using devtools:

devtools::install_github("metacran/cranlogs")

Now let’s load all the packages we will need:

library(mlr)
library(stringi)
library(cranlogs)
library(data.table)

Do obtain a neat table of all available learners in mlr we can call listLearners(). This table also contains a column with the needed packages for each learner separated with a ,.

# obtain used packages for all learners
lrns = as.data.table(listLearners())
all.pkgs = stri_split(lrns$package, fixed = ",")

Note: You might get some warnings here because you likely did not install all packages that mlr suggests – which is totally fine.

Now we can obtain the download counts from the rstudio cran mirror, i.e. from the last month. We use data.table to easily sum up the download counts of each day.

all.downloads = cran_downloads(packages = unique(unlist(all.pkgs)),
                               when = "last-month")
all.downloads = as.data.table(all.downloads)
monthly.downloads = all.downloads[, list(monthly = sum(count)), by = package]

As some learners need multiple packages we will use the download count of the package with the least downloads.

lrn.downloads = sapply(all.pkgs, function(pkgs) {
  monthly.downloads[package %in% pkgs, min(monthly)]
})

Let’s put these numbers in our table:

lrns$downloads = lrn.downloads
lrns = lrns[order(downloads, decreasing = TRUE),]
lrns[, .(class, name, package, downloads)]

Here are the first 5 rows of the table:

Now let’s get rid of the duplicates introduced by the distinction of the type classif, regr and we already have our…

Nearly final table

lrns.small = lrns[, .SD[1,], by = .(name, package)]
lrns.small[, .(class, name, package, downloads)]

The top 20 according to the rstudio cran mirror:

As we are just looking for the packages let’s compress the table a bit further and come to our…

Final table

lrns.pgks = lrns[,list(learners = paste(class, collapse = ",")),
                 by = .(package, downloads)]
lrns.pgks

Here are the first 20 rows of the table:

And of course we want to have a small visualization:

library(ggplot2)
library(forcats)
lrns.pgks$learners = factor(lrns.pgks$learners, lrns.pgks$learners)
g = ggplot(lrns.pgks[20:1], aes(x = fct_inorder(stri_sub(
  paste0(package,": ",learners), 0, 64)), y = downloads, fill = downloads))
g + geom_bar(stat = "identity") +
  coord_flip() +
  xlab("") +
  scale_fill_continuous(guide=FALSE)

Remarks

This is not really representative of how popular each learner is, as some packages have multiple purposes (e.g. multiple learners). Furthermore it would be great to have access to the trending list. Also most stars at GitHub gives a better view of what the developers are interested in. Looking for machine learning packages we see there e.g: xgboost, h2o and tensorflow.