Composable Preprocessing Operators and Pipelines for Machine Learning

Toolset that enriches 'mlr' with a diverse set of preprocessing operators. Composable Preprocessing Operators ("CPO"s) are first-class R objects that can be applied to data.frames and 'mlr' "Task"s to modify data, can be attached to 'mlr' "Learner"s to add preprocessing to machine learning algorithms, and can be composed to form preprocessing pipelines.

Build Status Coverage CRAN Status Badge CRAN Downloads

GSoC 2017 Project: Operator Based Machine Learning Pipeline Construction

What is CPO?

> task %<>>% cpoScale(scale = FALSE) %>>% cpoPca() %>>%  # pca
>   cpoFilterChiSquared(abs = 3) %>>%  # filter
>   cpoModelMatrix(~ 0 + .^2)  # interactions
> head(getTaskData(task))
        PC1        PC2         PC3    PC1:PC2     PC1:PC3      PC2:PC3 Species
1 -2.684126 -0.3193972  0.02791483  0.8573023 -0.07492690 -0.008915919  setosa
2 -2.714142  0.1770012  0.21046427 -0.4804064 -0.57122986  0.037252434  setosa
3 -2.888991  0.1449494 -0.01790026 -0.4187575  0.05171367 -0.002594632  setosa
4 -2.745343  0.3182990 -0.03155937 -0.8738398  0.08664130 -0.010045316  setosa
5 -2.728717 -0.3267545 -0.09007924  0.8916204  0.24580071  0.029433798  setosa
6 -2.280860 -0.7413304 -0.16867766  1.6908707  0.38473006  0.125045884  setosa

"Composable Preprocessing Operators" are an extension for the mlr ("Machine Learning in R") project which represent preprocessing operations (e.g. imputation or PCA) in the form of R objects. These CPO objects can be composed to form more complex operations, they can be applied to data sets, and they can be attached to mlr Learner objects to generate complex machine learning pipelines that perform both preprocessing and model fitting.

Table of Contents

Short Overview

CPOs are created by calling a constructor.

> cpoScale()
scale(center = TRUE, scale = TRUE)

The created objects have Hyperparameters that can be manipulated using getHyperPars, setHyperPars etc, just like in mlr.

> getHyperPars(cpoScale())
[1] TRUE
[1] TRUE
> setHyperPars(cpoScale(), = FALSE)
scale(center = FALSE, scale = TRUE)

The %>>%-operator can be used to create complex pipelines.

> cpoScale() %>>% cpoPca()
(scale >> pca)( = TRUE, scale.scale = TRUE)

This operator can also be used to apply an operation to a data set:

> head(iris %>>% cpoPca())
  Species       PC1      PC2          PC3          PC4
1  setosa -5.912747 2.302033  0.007401536  0.003087706
2  setosa -5.572482 1.971826  0.244592251  0.097552888
3  setosa -5.446977 2.095206  0.015029262  0.018013331
4  setosa -5.436459 1.870382  0.020504880 -0.078491501
5  setosa -5.875645 2.328290 -0.110338269 -0.060719326
6  setosa -6.477598 2.324650 -0.237202487 -0.021419633

Or to attach an operation to an MLR Learner, which extends the Learner's hyperparameters by the CPO's hyperparameters:

> cpoScale() %>>% makeLearner("classif.logreg")
Learner classif.logreg.scale from package stats
Type: classif
Name: ; Short name: 
Class: CPOLearner
Properties: numerics,factors,prob,twoclass
Predict-Type: response
Hyperparameters: model=FALSE,,scale.scale=TRUE

Get a list of all CPOs by calling listCPO().


Install mlrCPO from CRAN, or use the more recent GitHub version:



To effectively use mlrCPO, you should first familiarize yourself a little with mlr. There is an extensive tutorial online; for more ressources on mlr, see the overview on mlr's GitHub page.

To get familiar with mlrCPO, it is recommended that you read the vignettes. For each vignette, there is also a compact version that has all the R output removed.

  1. First Steps: Introduction and short overview (compact version).
  2. mlrCPO Core: Description of general tools for CPO handling (compact version).
  3. Builtin CPOs: Listing and description of all builtin CPOs (compact version).
  4. Custom CPOs: How to create your own CPOs. (compact version).
  5. CPO Internals: A small intro guide for developers into the code base. See the info directory for pdf / html versions.

For more documentation of individual mlrCPO functions, use R's built-in help() functionality.

Project Status

The foundation of mlrCPO is built and is reasonably stable, only small improvements and stability fixes are expected here. There are still many concrete implementations of preprocessing operators to be written.


Bugs, Questions, Feedback

mlrCPO is a free and open source software project that encourages participation and feedback. If you have any issues, questions, suggestions or feedback, please do not hesitate to open an "issue" about it on the GitHub page!

In case of problems / bugs, it is often helpful if you provide a "minimum working example" that showcases the behaviour (but don't worry about this if the bug is obvious).

Please understand that the ressources of the project are limited: response may sometimes be delayed by a few days, and some suggestions may not not make it to become features for a while.

Contributing Code, Pull Requests

Pull Requests that fix small issues are very welcome, especially if they contain tests that check for the given issue. For larger contributions, or Pull Requests that add features, please note:

  1. Adding new CPOs is always welcome. Please have a look at a few examples in the current codebase (the PCA CPO and the corresponding tests file are good for this, and show that adding a CPO does not require a lot of code) to familiarise yourself with the conventions. A CPO that comes with documentation, in particular also documenting the CPOTrained state, and with tests, is most likely to get merged quickly.

  2. Adding or changing features of the backend, or changing the functioning of the backend, is a more complicated story. If a Pull Request is incongruent with the "vision" behind mlrCPO, or if it appears to put a large burden on the mlrCPO developers in the long term relative to the problems it solves, it may have a slim chance of getting merged. Therefore, if you plan to make a contribution changing CPO core behaviour, it is best if you first open an "issue" about it for discussion.

When creating Pull Requests, please follow the Style Guide. Adherence to this is checked by the CI system (Travis). On Linux (and possibly Mac) you can check this locally on your computer using the quicklint tool in the tools directory. This is recommended to avoid frustrating failed builds caused by style violations.

Before merging a Pull Request, it is possible that an mlrCPO developer makes further changes to it, e.g. to harmonise it with conventions, or to incorporate other ideas.

When you make a Pull Request, it is assumed that you permit us (and are able to permit us) to incorporate the given code into the mlrCPO codebase as given, or with modifications, and distribute the result under the BSD 2-Clause License.

Similar Projects

There are other projects that provide functionality similar to mlrCPO for other machine learning frameworks. The caret project provides some preprocessing functionality, though not as flexible as mlrCPO. dplyr has similar syntax and some overlapping functionality, but is focused ultimately more on (manual) data manipulation instead of (machine learning pipeline integrated) preprocessing. Much more close to mlrCPO's functionality is the Recipes package. scikit learn also has preprocessing functionality built in.


The BSD 2-Clause License


mlrCPO 0.3.4

  • minor fix for DESCRIPTION file, required by CRAN. Functionally equivalent to 0.3.3

mlrCPO 0.3.3

  • minor fix for tests, required by CRAN

mlrCPO 0.3.2

  • minor bugfixes

mlrCPO 0.3.1

  • new cpo cpoMakeCols, cpoAddCols, cpoIca, cpoApplyFunRegrTarget, cpoLogTrafoRegr, cpoRegrResiduals, cpoSmote
  • refactoring of cpoPca, cpoApplyFun
  • Split vignettes up into 4 vignettes
  • Verbose printing using exclamation point: !cpo
  • slight change to makeCPO API

mlrCPO 0.3

  • great code and test quality improvement, bugfixes
  • retrafoless and target operating CPO backend works
  • new cpo cpoCache

mlrCPO 0.2

  • refactoring retrafoless and target operating CPOs (they still don't work)
  • standardised file names
  • much improved documentation
  • more coherent makeCPO* functions
  • %<>>% and %<<<% operators for compose / attach + assignment
  • %>|% and %|<% operators for apply + retrafo()
  • More coherent handling of cababilities
  • New getters: getCPOTrainedCapability, getOriginalCPO, getOriginalCPOConstructor
  • Inverter now can be inspected as / built from states
  • new property handling: adding.min and needed.max, for different standards internally & externally
  • new cpo cpoSelectFreeProperties, cpoCache, cpoTransformParams

mlrCPO 0.1

  • Initial release on GitHub, retrafoless and target operating CPOs don't work yet

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


0.3.7-3 by Martin Binder, 2 months ago

Report a bug at

Browse source code at

Authors: Martin Binder [aut, cre] , Bernd Bischl [ctb] , Michel Lang [ctb] , Lars Kotthoff [ctb]

Documentation:   PDF Manual  

BSD_2_clause + file LICENSE license

Imports BBmisc, stringi, checkmate, methods, stats, utils, backports

Depends on ParamHelpers, mlr

Suggests care, party, rpart, mlbench, knitr, rmarkdown, testthat, mRMRe, praznik, randomForestSRC, randomForest, ranger, Rfast, FSelector, FSelectorRcpp, e1071, DiscriMiner, digest, lintr, Hmisc, fastICA, rex

Imported by mosmafs.

Suggested by nntrf.

See at CRAN