Skip to content

Extension Packages

Michel Lang edited this page Oct 2, 2020 · 55 revisions

This page collects packages or projects which extend mlr3. Open an issue if you want your project listed here.

Overview

See this figure for an encompassing overview.

Released

Black-box optimization toolkit.

Extensions for cluster analysis.

Additional data sets (and tasks).

DataBackend for dbplyr. Allow tasks to transparently operate on various data storage systems, including SQL data bases, Apache Spark or Google's BigQuery.

Variable selection filters.

Variable selection wrappers like sequential forward/backward search, exhaustive search or genetic algorithms.

Hyperparameter tuning via hyperband.

Additional learners for regression and classification.

Many classification and regression performance measures, implemented as simple functions. Package comes with very few dependencies so it can easily be used in third-party packages.

Connector to OpenML.

Pipelines and DAGs for preprocessing and building complex workflows.

Extensions for supervised probabilistic learning (this includes survival analysis).

Hyperparameter tuning via random search, grid search, ...

Meta-package for installing and loading released core packages.

Visualization via ggplot2's autoplot() function.

In Development

Connector between mlr3 and batchtools

Extensions for time-series forecasting.

Extensions for deep learning via keras.

Hyperparameter tuning via model-based optimization (a.k.a. Bayesian optimization).

Extensions for ordinal regression.

Extensions for resampling spatio-temporal tasks.

Planned

Open an issue in the respective project if you are interested in starting or supporting the development.

Extension to source and manage additional learners from the mlr3learners organization on GitHub.

Extensions for functional data analysis.

Extensions for multilabel classification.