This repo contains material for a workshop that teaches how to create ensemble models in R using the mlr package. Click on ensembling_workshop.md above to see the output of the tutorial. If you would like to pull down the repo and run the algorithms yourself, use ensembling_workshop.Rmd.
You will need to install several packages in order to run the code:
install.packages("devtools")
install.packages("mlr", dependencies = TRUE, suggests = TRUE)
install.packages("mlrMBO", dependencies = TRUE)
install.packages("tidyverse", dependencies = TRUE)
devtools::install_github("Prometheus77/ucimlr")
devtools::install_github("Prometheus77/actools")
Installing the mlr packages with all suggested dependences can take upwards of an hour. As an alternative, you can remove the suggests = TRUE code and then install packages as you go. Every time mlr needs a package you don't have installed, it will throw an error and tell you which package is missing.
- mlr tutorial: https://mlr.mlr-org.com/
- mlr Github repo: https://github.com/mlr-org/mlr
- mlr cheat sheet: https://github.com/mlr-org/mlr/blob/master/addon/cheatsheet/MlrCheatsheet.pdf
- Ensemble selection white paper: http://www.cs.cornell.edu/~alexn/papers/shotgun.icml04.revised.rev2.pdf