Mlr

The mlr package is a machine learning framework for the R programming language. It provides a single interface to a large number of classification, regression, survival analysis and clustering methods, so users do not need to write separate wrappers for each algorithm in R. Around that interface it supplies the surrounding infrastructure for experiments: generic resampling such as cross validation, bootstrapping and subsampling, hyperparameter tuning with modern optimization techniques for single and multi objective problems, filter and wrapper methods for feature selection, preprocessing and imputation, nested resampling, benchmarking to compare models, performance measures and visualization, and parallelization of most operations. Users can extend it with custom learners, measures, imputation methods and filters. The site hosts an online tutorial, a cheat sheet and a changelog. It is aimed at R users, statisticians and researchers running machine learning experiments. The mlr team has retired the package. No new features will be added and only severe bugs will be fixed, and the maintainers recommend its successor framework, mlr3, for new projects. The package is installed from CRAN or from its GitHub repository, and the site publishes no pricing.