Function reference • mlr3pipelines

Package

mlr3pipelines mlr3pipelines-package: mlr3pipelines: Preprocessing Operators and Pipelines for 'mlr3'

Building Blocks

PipeOp: PipeOp Base Class

Graph: Graph Base Class

PipeOpTaskPreproc: Task Preprocessing Base Class

PipeOpTaskPreprocSimple: Simple Task Preprocessing Base Class

PipeOpImpute: Imputation Base Class

Multiplicity(): Multiplicity

Graph Tools

`%>>%` concat_graphs() `%>>!%`: PipeOp Composition Operator

gunion(): Disjoint Union of Graphs

greplicate(): Create Disjoint Graph Union of Copies of a Graph

PipeOps

mlr_pipeops: Dictionary of PipeOps

po() pos(): Shorthand PipeOp Constructor

mlr_pipeops_boxcox PipeOpBoxCox: Box-Cox Transformation of Numeric Features

mlr_pipeops_branch PipeOpBranch: Path Branching

mlr_pipeops_chunk PipeOpChunk: Chunk Input into Multiple Outputs

mlr_pipeops_classbalancing PipeOpClassBalancing: Class Balancing

mlr_pipeops_classifavg PipeOpClassifAvg: Majority Vote Prediction

mlr_pipeops_classweights PipeOpClassWeights: Class Weights for Sample Weighting

mlr_pipeops_colapply PipeOpColApply: Apply a Function to each Column of a Task

mlr_pipeops_collapsefactors PipeOpCollapseFactors: Collapse Factors

mlr_pipeops_colroles PipeOpColRoles: Change Column Roles of a Task

mlr_pipeops_copy PipeOpCopy: Copy Input Multiple Times

mlr_pipeops_datefeatures PipeOpDateFeatures: Preprocess Date Features

mlr_pipeops_encode PipeOpEncode: Factor Encoding

mlr_pipeops_encodeimpact PipeOpEncodeImpact: Conditional Target Value Impact Encoding

mlr_pipeops_encodelmer PipeOpEncodeLmer: Impact Encoding with Random Intercept Models

mlr_pipeops_featureunion PipeOpFeatureUnion: Aggregate Features from Multiple Inputs

mlr_pipeops_filter PipeOpFilter: Feature Filtering

mlr_pipeops_fixfactors PipeOpFixFactors: Fix Factor Levels

mlr_pipeops_histbin PipeOpHistBin: Split Numeric Features into Equally Spaced Bins

mlr_pipeops_ica PipeOpICA: Independent Component Analysis

mlr_pipeops_imputeconstant PipeOpImputeConstant: Impute Features by a Constant

mlr_pipeops_imputehist PipeOpImputeHist: Impute Numerical Features by Histogram

mlr_pipeops_imputelearner PipeOpImputeLearner: Impute Features by Fitting a Learner

mlr_pipeops_imputemean PipeOpImputeMean: Impute Numerical Features by their Mean

mlr_pipeops_imputemedian PipeOpImputeMedian: Impute Numerical Features by their Median

mlr_pipeops_imputemode PipeOpImputeMode: Impute Features by their Mode

mlr_pipeops_imputeoor PipeOpImputeOOR: Out of Range Imputation

mlr_pipeops_imputesample PipeOpImputeSample: Impute Features by Sampling

mlr_pipeops_kernelpca PipeOpKernelPCA: Kernelized Principle Component Analysis

mlr_pipeops_learner PipeOpLearner: Wrap a Learner into a PipeOp

mlr_pipeops_learner_cv PipeOpLearnerCV: Wrap a Learner into a PipeOp with Cross-validated Predictions as Features

mlr_pipeops_missind PipeOpMissInd: Add Missing Indicator Columns

mlr_pipeops_modelmatrix PipeOpModelMatrix: Transform Columns by Constructing a Model Matrix

mlr_pipeops_multiplicityexply PipeOpMultiplicityExply: Explicate a Multiplicity

mlr_pipeops_multiplicityimply PipeOpMultiplicityImply: Implicate a Multiplicity

mlr_pipeops_mutate PipeOpMutate: Add Features According to Expressions

mlr_pipeops_nmf PipeOpNMF: Non-negative Matrix Factorization

mlr_pipeops_nop PipeOpNOP: Simply Push Input Forward

mlr_pipeops_ovrsplit PipeOpOVRSplit: Split a Classification Task into Binary Classification Tasks

mlr_pipeops_ovrunite PipeOpOVRUnite: Unite Binary Classification Tasks

mlr_pipeops_pca PipeOpPCA: Principle Component Analysis

mlr_pipeops_proxy PipeOpProxy: Wrap another PipeOp or Graph as a Hyperparameter

mlr_pipeops_quantilebin PipeOpQuantileBin: Split Numeric Features into Quantile Bins

mlr_pipeops_randomprojection PipeOpRandomProjection: Project Numeric Features onto a Randomly Sampled Subspace

mlr_pipeops_randomresponse PipeOpRandomResponse: Generate a Randomized Response Prediction

mlr_pipeops_regravg PipeOpRegrAvg: Weighted Prediction Averaging

mlr_pipeops_removeconstants PipeOpRemoveConstants: Remove Constant Features

mlr_pipeops_renamecolumns PipeOpRenameColumns: Rename Columns

mlr_pipeops_replicate PipeOpReplicate: Replicate the Input as a Multiplicity

mlr_pipeops_scale PipeOpScale: Center and Scale Numeric Features

mlr_pipeops_scalemaxabs PipeOpScaleMaxAbs: Scale Numeric Features with Respect to their Maximum Absolute Value

mlr_pipeops_scalerange PipeOpScaleRange: Linearly Transform Numeric Features to Match Given Boundaries

mlr_pipeops_select PipeOpSelect: Remove Features Depending on a Selector

mlr_pipeops_smote PipeOpSmote: SMOTE Balancing

mlr_pipeops_spatialsign PipeOpSpatialSign: Normalize Data Row-wise

mlr_pipeops_subsample PipeOpSubsample: Subsampling

mlr_pipeops_targetinvert PipeOpTargetInvert: Invert Target Transformations

mlr_pipeops_targetmutate PipeOpTargetMutate: Transform a Target by a Function

mlr_pipeops_targettrafoscalerange PipeOpTargetTrafoScaleRange: Linearly Transform a Numeric Target to Match Given Boundaries

mlr_pipeops_textvectorizer PipeOpTextVectorizer: Bag-of-word Representation of Character Features

mlr_pipeops_threshold PipeOpThreshold: Change the Threshold of a Classification Prediction

mlr_pipeops_tunethreshold PipeOpTuneThreshold: Tune the Threshold of a Classification Prediction

mlr_pipeops_unbranch PipeOpUnbranch: Unbranch Different Paths

mlr_pipeops_updatetarget PipeOpUpdateTarget: Transform a Target without an Explicit Inversion

mlr_pipeops_vtreat PipeOpVtreat: Interface to the vtreat Package

mlr_pipeops_yeojohnson PipeOpYeoJohnson: Yeo-Johnson Transformation of Numeric Features

PipeOp: PipeOp Base Class

PipeOpEnsemble: Ensembling Base Class

PipeOpImpute: Imputation Base Class

PipeOpTargetTrafo: Target Transformation Base Class

PipeOpTaskPreproc: Task Preprocessing Base Class

PipeOpTaskPreprocSimple: Simple Task Preprocessing Base Class

Pipelines

ppl() ppls(): Shorthand Graph Constructor

pipeline_bagging(): Create a bagging learner

pipeline_branch(): Branch Between Alternative Paths

pipeline_greplicate(): Create Disjoint Graph Union of Copies of a Graph

pipeline_ovr(): Create A Graph to Perform "One vs. Rest" classification.

pipeline_robustify(): Robustify a learner

pipeline_stacking(): Create A Graph to Perform Stacking.

pipeline_targettrafo(): Transform and Re-Transform the Target Variable

Graphs

mlr_graphs: Dictionary of (sub-)graphs

pipeline_bagging(): Create a bagging learner

pipeline_branch(): Branch Between Alternative Paths

pipeline_greplicate(): Create Disjoint Graph Union of Copies of a Graph

pipeline_ovr(): Create A Graph to Perform "One vs. Rest" classification.

pipeline_robustify(): Robustify a learner

pipeline_stacking(): Create A Graph to Perform Stacking.

pipeline_targettrafo(): Transform and Re-Transform the Target Variable

chain_graphs(): Chain a Series of Graphs

Learners

mlr_learners_graph GraphLearner: Encapsulate a Graph as a Learner

mlr_learners_classif.avg mlr_learners_regr.avg: Optimized Weighted Average of Features for Classification and Regression

Helpers

selector_all() selector_none() selector_type() selector_grep() selector_name() selector_invert() selector_intersect() selector_union() selector_setdiff() selector_missing() selector_cardinality_greater_than(): Selector Functions

as_graph(): Conversion to mlr3pipelines Graph

assert_graph(): Assertion for mlr3pipelines Graph

as_pipeop(): Conversion to mlr3pipelines PipeOp

assert_pipeop(): Assertion for mlr3pipelines PipeOp

is_noop(): Test for NO_OP

NO_OP: No-Op Sentinel Used for Alternative Branching

filter_noop(): Remove NO_OPs from a List

as.Multiplicity(): Convert an object to a Multiplicity

is.Multiplicity(): Check if an object is a Multiplicity

Abstract PipeOps

PipeOpEnsemble: Ensembling Base Class

PipeOp Type Inference

add_class_hierarchy_cache(): Add a Class Hierarchy to the Cache

reset_class_hierarchy_cache(): Reset the Class Hierarchy Cache

register_autoconvert_function(): Add Autoconvert Function to Conversion Register

reset_autoconvert_register(): Reset Autoconvert Register