Combines a predicted reponse and se from PredictionRegr with a specified probability distribution to estimate (or 'compose') a distr prediction.

## Dictionary

This PipeOp can be instantiated via the dictionary mlr3pipelines::mlr_pipeops or with the associated sugar function mlr3pipelines::po():

PipeOpProbregrCompositor$new() mlr_pipeops$get("compose_probregr")
po("compose_probregr")

## Input and Output Channels

PipeOpProbregrCompositor has two input channels named "input_response" and "input_se", which take NULL during training and two PredictionRegrs during prediction, these should respectively contain the response and se return type, the same object can be passed twice.

The output during prediction is a PredictionRegr with the "response" from input_response, the "se" from input_se and a "distr" created from combining the two.

## State

The $state is left empty (list()). ## Parameters • dist :: character(1) Location-scale distribution to use for composition. Current choices are "Normal" (default), "Cauchy", "Gumbel", "Laplace", "Logistic". All implemented via distr6. ## Internals The composition is created by substituting the response and se predictions into the distribution location and scale parameters respectively. ## Super class mlr3pipelines::PipeOp -> PipeOpProbregrCompositor ## Methods ### Public methods Inherited methods ### Method new() Creates a new instance of this R6 class. #### Usage PipeOpProbregrCompositor$new(
id = "compose_probregr",
param_vals = list(dist = "Normal")
)

#### Arguments

id

(character(1))
Identifier of the resulting object.

param_vals

(list())
List of hyperparameter settings, overwriting the hyperparameter settings that would otherwise be set during construction.

### Method clone()

The objects of this class are cloneable with this method.

PipeOpProbregrCompositor$clone(deep = FALSE) #### Arguments deep Whether to make a deep clone. ## Examples if (FALSE) { if (requireNamespace("mlr3pipelines", quietly = TRUE) && requireNamespace("rpart", quietly = TRUE)) { library(mlr3) library(mlr3pipelines) set.seed(1) task = tsk("boston_housing") # Option 1: Use a learner that can predict se learn = lrn("regr.featureless", predict_type = "se") pred = learn$train(task)$predict(task) poc = po("compose_probregr") poc$predict(list(pred, pred))[[1]]

# Option 2: Use two learners, one for response and the other for se
learn_response = lrn("regr.rpart")
learn_se = lrn("regr.featureless", predict_type = "se")
pred_response = learn_response$train(task)$predict(task)
pred_se = learn_se$train(task)$predict(task)
poc = po("compose_probregr")
poc\$predict(list(pred_response, pred_se))[[1]]
}
}