Classification via regression wrapper.

Builds regression models that predict for the positive class whether a particular example belongs to it (1) or not (-1).

Probabilities are generated by transforming the predictions with a softmax.

Inspired by WEKA's ClassificationViaRegression (http://weka.sourceforge.net/doc.dev/weka/classifiers/meta/ClassificationViaRegression.html).

Usage

makeClassificationViaRegressionWrapper(learner, predict.type = "response")

Arguments

learner

(Learner | character(1))
The learner. If you pass a string the learner will be created via makeLearner.

predict.type

(character(1))
“response” (= labels) or “prob” (= probabilities and labels by selecting the one with maximal probability).

Value

Learner.

See also

Other wrapper: makeBaggingWrapper(), makeConstantClassWrapper(), makeCostSensClassifWrapper(), makeCostSensRegrWrapper(), makeDownsampleWrapper(), makeDummyFeaturesWrapper(), makeExtractFDAFeatsWrapper(), makeFeatSelWrapper(), makeFilterWrapper(), makeImputeWrapper(), makeMulticlassWrapper(), makeMultilabelBinaryRelevanceWrapper(), makeMultilabelClassifierChainsWrapper(), makeMultilabelDBRWrapper(), makeMultilabelNestedStackingWrapper(), makeMultilabelStackingWrapper(), makeOverBaggingWrapper(), makePreprocWrapperCaret(), makePreprocWrapper(), makeRemoveConstantFeaturesWrapper(), makeSMOTEWrapper(), makeTuneWrapper(), makeUndersampleWrapper(), makeWeightedClassesWrapper()

Examples

lrn = makeLearner("regr.rpart")
lrn = makeClassificationViaRegressionWrapper(lrn)
mod = train(lrn, sonar.task, subset = 1:140)
#> Error in x[0, , drop = FALSE]: incorrect number of dimensions
predictions = predict(mod, newdata = getTaskData(sonar.task)[141:208, 1:60])
#> Error in predict(mod, newdata = getTaskData(sonar.task)[141:208, 1:60]): object 'mod' not found