Skip to contents

Gradient Boosting Classification Algorithm. Calls gbm::gbm() from gbm.

Dictionary

This Learner can be instantiated via lrn():

lrn("classif.gbm")

Meta Information

  • Task type: “classif”

  • Predict Types: “response”, “prob”

  • Feature Types: “integer”, “numeric”, “factor”, “ordered”

  • Required Packages: mlr3, mlr3extralearners, gbm

Parameters

IdTypeDefaultLevelsRange
distributioncharacterbernoullibernoulli, adaboost, huberized, multinomial-
n.treesinteger100\([1, \infty)\)
interaction.depthinteger1\([1, \infty)\)
n.minobsinnodeinteger10\([1, \infty)\)
shrinkagenumeric0.001\([0, \infty)\)
bag.fractionnumeric0.5\([0, 1]\)
train.fractionnumeric1\([0, 1]\)
cv.foldsinteger0\((-\infty, \infty)\)
keep.datalogicalFALSETRUE, FALSE-
verboselogicalFALSETRUE, FALSE-
n.coresinteger1\((-\infty, \infty)\)
var.monotoneuntyped--

Initial parameter values

  • keep.data is initialized to FALSE to save memory.

  • n.cores is initialized to 1 to avoid conflicts with parallelization through future.

References

Friedman, H J (2002). “Stochastic gradient boosting.” Computational statistics & data analysis, 38(4), 367–378.

See also

Author

be-marc

Super classes

mlr3::Learner -> mlr3::LearnerClassif -> LearnerClassifGBM

Methods

Inherited methods


Method new()

Creates a new instance of this R6 class.

Usage


Method importance()

The importance scores are extracted by gbm::relative.influence() from the model.

Usage

LearnerClassifGBM$importance()

Returns

Named numeric().


Method clone()

The objects of this class are cloneable with this method.

Usage

LearnerClassifGBM$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# Define the Learner
learner = mlr3::lrn("classif.gbm")
print(learner)
#> <LearnerClassifGBM:classif.gbm>: Gradient Boosting
#> * Model: -
#> * Parameters: keep.data=FALSE, n.cores=1
#> * Packages: mlr3, mlr3extralearners, gbm
#> * Predict Types:  [response], prob
#> * Feature Types: integer, numeric, factor, ordered
#> * Properties: importance, missings, twoclass, weights

# Define a Task
task = mlr3::tsk("sonar")

# Create train and test set
ids = mlr3::partition(task)

# Train the learner on the training ids
learner$train(task, row_ids = ids$train)
#> Distribution not specified, assuming bernoulli ...

print(learner$model)
#> gbm::gbm(formula = f, data = data, keep.data = FALSE, n.cores = 1L)
#> A gradient boosted model with bernoulli loss function.
#> 100 iterations were performed.
#> There were 60 predictors of which 42 had non-zero influence.
print(learner$importance())
#>        V11         V9        V12        V44        V13        V31        V51 
#> 12.6008956 11.7525462 11.6497257  7.0666897  6.7942342  6.1490541  6.0363449 
#>        V22        V52        V36        V49        V48        V40        V39 
#>  4.5755311  4.5120352  4.4656788  4.4634687  4.1842278  3.9818023  3.2945606 
#>        V37        V27        V10        V54        V32         V4        V45 
#>  3.2614260  3.1027780  2.9733663  2.7780510  2.6653824  2.4417863  2.3918495 
#>        V16        V60        V26        V43        V46        V47        V55 
#>  2.3411173  2.2633236  1.5173297  1.2631283  1.2240094  1.1395545  1.1339056 
#>        V53        V15         V1        V21        V59        V42        V17 
#>  1.0637253  1.0300141  0.9857307  0.9020723  0.6087593  0.5901937  0.5892742 
#>        V23        V20        V56         V8        V38        V50        V19 
#>  0.5739356  0.5071219  0.4772756  0.4739872  0.4502871  0.4286625  0.4138723 
#>        V14        V18         V2        V24        V25        V28        V29 
#>  0.0000000  0.0000000  0.0000000  0.0000000  0.0000000  0.0000000  0.0000000 
#>         V3        V30        V33        V34        V35        V41         V5 
#>  0.0000000  0.0000000  0.0000000  0.0000000  0.0000000  0.0000000  0.0000000 
#>        V57        V58         V6         V7 
#>  0.0000000  0.0000000  0.0000000  0.0000000 

# Make predictions for the test rows
predictions = learner$predict(task, row_ids = ids$test)

# Score the predictions
predictions$score()
#> classif.ce 
#>   0.173913