mlr3 (version 0.1.1)

LearnerClassifDebug: Classification Learner for Debugging

Description

A simple LearnerClassif used primarily in the unit tests and for debugging purposes. If no hyperparameter is set, it simply constantly predicts a randomly selected label. The following hyperparameters trigger the following actions:

message_train:

Outputs a message during train if runif(1) exceeds its value.

message_predict:

Outputs a message during predict if runif(1) exceeds its value.

warning_train:

Signals a warning during train if runif(1) exceeds its value.

warning_predict:

Signals a warning during predict if runif(1) exceeds its value.

error_train:

Raises an exception during train if runif(1) exceeds its value.

error_predict:

Raises an exception during predict if runif(1) exceeds its value.

segfault_train:

Provokes a segfault during train if runif(1) exceeds its value.

segfault_predict:

Provokes a segfault during predict if runif(1) exceeds its value.

predict_missing

Ratio of predictions which will be NA.

save_tasks:

Saves input task in model slot during training and prediction.

x:

Numeric parameter. Ignored.

Note that segfaults may not work on your operating system. Also note that if they work, they will tear down your R session immediately!

Usage

LearnerClassifDebug

Format

R6::R6Class inheriting from LearnerClassif.

Examples

# NOT RUN {
learner = LearnerClassifDebug$new()
learner$param_set$values = list(message_train = 1, save_tasks = TRUE)

# this should signal a message
task = mlr_tasks$get("iris")
learner$train(task)
learner$predict(task)

# task_train and task_predict are the input tasks for train() and predict()
names(learner$model)
# }