Configures the behavior of the package.
Failure model.
Prediction from resampling.
Estimate the residual variance
Aggregation methods.
Set, add, remove or query properties of learners
Predict new data with an R learner.
Create a feature filter
Subset data in task.
Plot threshold vs. performance(s) for 2-class classification.
Tune prediction threshold.
Iris cost-sensitive classification task
getCostSensWeightedPairsModels
Returns the list of fitted models.
Show and visualize the steps of feature selection.
Is the model a FailureModel?
Predict new data.
Instantiates a resampling strategy object.
List of supported learning algorithms.
Motor Trend Car Road Tests clustering task
Normalize features
Join some class existing levels to new, larger class levels for classification problems.
makeCustomResampledMeasure
Construct your own resampled performance measure.
Create a stacked learner object.
Fuse learner with simple ove/underrsampling for imbalancy correction in binary classification.
Return task ids used in benchmark.
Returns the list of models fitted in bagging.
Returns the underlying classification model.
Visualizes a learning algorithm on a 1D or 2D data set.
Result of multi-criteria tuning.
makeWeightedClassesWrapper
Wraps a classifier for weighted fitting where each class receives a weight.
Result of tuning.
Wraps a regression learner for use in cost-sensitive learning.
Prediction object.
An aggregation method reduces the performance values of the test
(and possibly the training sets) to a single value.
To see all possible, implemented aggregations look at aggregations
. Fuse learner with tuning.
Description object for task.
European Union Agricultural Workforces clustering task
crossover
Set the id of a learner object.
Get target column of task.
Converts predictions to a format package ROCR can handle.
makeTuneMultiCritControlGrid
Create control structures for multi-criteria tuning.
Fuse learner with the bagging technique and oversampling for imbalancy correction.
Train an R learner.
Specifiy your own aggregation of measures
Benchmark experiment for multiple learners and tasks.
Downsample (subsample) a task or a data.frame.
List filter methods
Wisconsin Breast Cancer classification task
Remove hyperparameters settings of a learner.
Create a description object for a resampling strategy.
Find matching learning algorithms.
Over- or undersample binary classification task to handle class imbalancy.
Wisonsin Prognostic Breast Cancer (WPBC) survival task
Internal construction / wrapping of learner object.
Display all possible hyperparameter settings for a learner that mlr knows.
Construct performance measure.
Summarize columns of data.frame or task.
Create a classification, regression, survival, cluster, or cost-sensitive classification task.
Remove constant features from a data set.
Measure performance of prediction.
Extract the aggregated performance values from a benchmark result.
Result of feature selection.
Synthetic Minority Oversampling Technique to handle class imbalancy in binary classification.
PimaIndiansDiabetes classification task
Returns a list of mlr's options
Fit models according to a resampling strategy.
Generate dummy variables for factor features.
Feature selection by wrapper approach.
Get underlying R model of learner integrated into mlr.
Returns the filtered features.
Plot filter values.
Set threshold of prediction object.
Sonar classification task
Get formula of a task.
Returns the list of fitted models.
Get feature names of task.
Fuse learner with the bagging technique.
Extract the feature selection results from a benchmark result.
makeFeatSelControlExhaustive
Create control structures for feature selection.
Summarizes factors of a data.frame by tabling them.
Re-impute a data set
Convert arguments to control structure.
Fuse learner with multiclass method.
Create a custom imputation method.
Extract the test performance values from a benchmark result.
Create control structures for tuning.
Induced model of learner.
makeCostSensClassifWrapper
Wraps a classification learner for use in cost-sensitive learning.
Plots multi-criteria results after tuning.
Find matching measures.
Result of a benchmark run.
Creates a measure for non-standard misclassification costs.
Set the type of predictions the learner should return.
getStackedBaseLearnerPredictions
Returns the predictions for each base learner.
Fuse learner with SMOTE oversampling for imbalancy correction in binary classification.
Get number of features in task.
Hyperparameter tuning for multiple measures at once.
Hyperparameter tuning.
Extract the feature selection results from a benchmark result.
Boston Housing regression task
Generate a fixed holdout instance for resampling.
Fuse learner with a feature filter method.
Create learner object.
Performance measures.
Get probabilities for some classes.
Impute and re-impute data
Set the hyperparameters of a learner object.
Fuse learner with preprocessing.
Merges small levels of factors into new level.
Fuse learner with an imputation method.
Create model multiplexer for model selection to tune over multiple possible models.
Train a learning algorithm.
Set aggregation function of measure.
Only exported for internal use.
Returns the selected feature set and optimization path after training.
Get current parameter settings for a learner.
Filter features by thresholding filter values.
Extract the predictions from a benchmark result.
Confusion matrix.
Extract the tuning results from a benchmark result.
Convert large/infinite numeric values in a data.frame or task.
Return learner ids used in benchmark.
Drop some features of task.
Returns the optimal hyperparameters and optimization path after training.
Iris classification task
Extract data in task.
NCCTG Lung Cancer survival task
Extract costs in task.
Get a description of all possible parameter settings for a learner.
Built in imputation methods
The built-ins are:
imputeConstant(const)
for imputation using a constant value,imputeMedian()
for imputation using the median,imputeMode()
for imputation using the mode,imputeMin(multiplier)
for imputing constant values shifted below the minimum
usingmin(x) - multiplier * diff(range(x))
,imputeMin(multiplier)
for imputing constant values shifted above the maximum
usingmax(x) + multiplier * diff(range(x))
,imputeNormal(mean, sd)
for imputation using normally
distributed random values. Mean and standard deviation will be calculated
from the data if not provided.imputeHist(breaks, use.mids)
for imputation using random values
with probabilities calculated usingtable
orhist
.imputeLearner(learner, preimpute)
for imputations using the response
of a classification or regression learner. Return error message of FailureModel.
makeModelMultiplexerParamSet
Creates a parameter set for model multiplexer tuning.
Fuse learner with simple downsampling (subsampling).
Fuse learner with feature selection.
makeCostSensWeightedPairsWrapper
Wraps a classifier for cost-sensitive learning to produce a weighted pairs model.
Calculates feature filter values.