Classification Imbalanced Random Forest Src Learner

Imbalanced Random forest for classification between two classes. Calls randomForestSRC::imbalanced.rfsrc() from from randomForestSRC.

Dictionary

This Learner can be instantiated via lrn():

lrn("classif.imbalanced_rfsrc")

Meta Information

• Task type: “classif”

• Predict Types: “response”, “prob”

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

• Required Packages: mlr3, randomForestSRC

Parameters

Id	Type	Default	Levels	Range
ntree	integer	500		\([1, \infty)\)
method	character	rfq	rfq, brf, standard	-
block.size	integer	10		\([1, \infty)\)
fast	logical	FALSE	TRUE, FALSE	-
ratio	numeric	-		\([0, 1]\)
mtry	integer	-		\([1, \infty)\)
mtry.ratio	numeric	-		\([0, 1]\)
nodesize	integer	15		\([1, \infty)\)
nodedepth	integer	-		\([1, \infty)\)
splitrule	character	gini	gini, auc, entropy	-
nsplit	integer	10		\([0, \infty)\)
importance	character	FALSE	FALSE, TRUE, none, permute, random, anti	-
bootstrap	character	by.root	by.root, by.node, none, by.user	-
samptype	character	swor	swor, swr	-
samp	untyped	-		-
membership	logical	FALSE	TRUE, FALSE	-
sampsize	untyped	-		-
sampsize.ratio	numeric	-		\([0, 1]\)
na.action	character	na.omit	na.omit, na.impute	-
nimpute	integer	1		\([1, \infty)\)
ntime	integer	-		\([1, \infty)\)
proximity	character	FALSE	FALSE, TRUE, inbag, oob, all	-
distance	character	FALSE	FALSE, TRUE, inbag, oob, all	-
forest.wt	character	FALSE	FALSE, TRUE, inbag, oob, all	-
xvar.wt	untyped	-		-
split.wt	untyped	-		-
forest	logical	TRUE	TRUE, FALSE	-
var.used	character	FALSE	FALSE, all.trees	-
split.depth	character	FALSE	FALSE, all.trees	-
seed	integer	-		\((-\infty, -1]\)
do.trace	logical	FALSE	TRUE, FALSE	-
get.tree	untyped	-		-
outcome	character	train	train, test	-
ptn.count	integer	0		\([0, \infty)\)
cores	integer	1		\([1, \infty)\)
save.memory	logical	FALSE	TRUE, FALSE	-
perf.type	character	-	gmean, misclass, brier, none	-
case.depth	logical	FALSE	TRUE, FALSE	-
marginal.xvar	untyped	NULL		-

Custom mlr3 parameters

• mtry: This hyperparameter can alternatively be set via the added hyperparameter mtry.ratio as mtry = max(ceiling(mtry.ratio * n_features), 1). Note that mtry and mtry.ratio are mutually exclusive.

• sampsize: This hyperparameter can alternatively be set via the added hyperparameter sampsize.ratio as sampsize = max(ceiling(sampsize.ratio * n_obs), 1). Note that sampsize and sampsize.ratio are mutually exclusive.

• cores: This value is set as the option rf.cores during training and is set to 1 by default.

References

O’Brien R, Ishwaran H (2019). “A random forests quantile classifier for class imbalanced data.” Pattern Recognition, 90, 232–249. doi:10.1016/j.patcog.2019.01.036 .

Chao C, Leo B (2004). “Using Random Forest to Learn Imbalanced Data.” University of California, Berkeley.

See also

• Dictionary of Learners: mlr3::mlr_learners.

• as.data.table(mlr_learners) for a table of available Learners in the running session (depending on the loaded packages).

• Chapter in the mlr3book: https://mlr3book.mlr-org.com/chapters/chapter2/data_and_basic_modeling.html#sec-learners

• mlr3learners for a selection of recommended learners.

• mlr3cluster for unsupervised clustering learners.

• mlr3pipelines to combine learners with pre- and postprocessing steps.

• mlr3tuning for tuning of hyperparameters, mlr3tuningspaces for established default tuning spaces.

Author

HarutyunyanLiana

Super classes

mlr3::Learner -> mlr3::LearnerClassif -> LearnerClassifImbalancedRandomForestSRC

Methods

Public methods

• LearnerClassifImbalancedRandomForestSRC$new()

• LearnerClassifImbalancedRandomForestSRC$importance()

• LearnerClassifImbalancedRandomForestSRC$selected_features()

• LearnerClassifImbalancedRandomForestSRC$oob_error()

• LearnerClassifImbalancedRandomForestSRC$clone()

Inherited methods

• mlr3::Learner$base_learner()
• mlr3::Learner$configure()
• mlr3::Learner$encapsulate()
• mlr3::Learner$format()
• mlr3::Learner$help()
• mlr3::Learner$predict()
• mlr3::Learner$predict_newdata()
• mlr3::Learner$print()
• mlr3::Learner$reset()
• mlr3::Learner$train()
• mlr3::LearnerClassif$predict_newdata_fast()

---

Method new()

Creates a new instance of this R6 class.

Usage

LearnerClassifImbalancedRandomForestSRC$new()

---

Method importance()

The importance scores are extracted from the slot importance.

Usage

LearnerClassifImbalancedRandomForestSRC$importance()

Returns

Named numeric().

---

Method selected_features()

Selected features are extracted from the model slot var.used.

Note: Due to a known issue in randomForestSRC, enabling var.used = "all.trees" causes prediction to fail. Therefore, this setting should be used exclusively for feature selection purposes and not when prediction is required.

Usage

LearnerClassifImbalancedRandomForestSRC$selected_features()

Returns

character().

---

Method oob_error()

OOB error extracted from the model slot err.rate.

Usage

LearnerClassifImbalancedRandomForestSRC$oob_error()

Returns

numeric().

---

Method clone()

The objects of this class are cloneable with this method.

Usage

LearnerClassifImbalancedRandomForestSRC$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# Define the Learner
learner = lrn("classif.imbalanced_rfsrc", importance = "TRUE")
print(learner)
#> 
#> ── <LearnerClassifImbalancedRandomForestSRC> (classif.imbalanced_rfsrc): Imbalan
#> • Model: -
#> • Parameters: importance=TRUE
#> • Packages: mlr3 and randomForestSRC
#> • Predict Types: [response] and prob
#> • Feature Types: logical, integer, numeric, factor, and ordered
#> • Encapsulation: none (fallback: -)
#> • Properties: importance, missings, oob_error, selected_features, twoclass, and
#> weights
#> • Other settings: use_weights = 'use', predict_raw = 'FALSE'

# Define a Task
task = tsk("sonar")
# Create train and test set
ids = partition(task)

# Train the learner on the training ids
learner$train(task, row_ids = ids$train)

print(learner$model)
#>                          Sample size: 139
#>            Frequency of class labels: M=75, R=64
#>                      Number of trees: 3000
#>            Forest terminal node size: 1
#>        Average no. of terminal nodes: 17.3687
#> No. of variables tried at each split: 8
#>               Total no. of variables: 60
#>        Resampling used to grow trees: swor
#>     Resample size used to grow trees: 88
#>                             Analysis: RFQ
#>                               Family: class
#>                       Splitting rule: auc *random*
#>        Number of random split points: 10
#>                     Imbalanced ratio: 1.1719
#>                    (OOB) Brier score: 0.14592221
#>         (OOB) Normalized Brier score: 0.58368883
#>                            (OOB) AUC: 0.89291667
#>                       (OOB) Log-loss: 0.4546593
#>                         (OOB) PR-AUC: 0.88827429
#>                         (OOB) G-mean: 0.80389676
#>    (OOB) Requested performance error: 0.19610324
#> 
#> Confusion matrix:
#> 
#>           predicted
#>   observed  M  R class.error
#>          M 66  9      0.1200
#>          R 17 47      0.2656
#> 
#>       (OOB) Misclassification rate: 0.1870504
#> 
#> Random-classifier baselines (uniform):
#>    Brier: 0.25   Normalized Brier: 1   Log-loss: 0.69314718
print(learner$importance())
#>          V12          V37          V11          V36           V9          V16 
#>  0.029300090  0.018482856  0.014646060  0.012273954  0.012273954  0.008598074 
#>          V17          V21          V27          V52           V1          V10 
#>  0.008598074  0.008598074  0.008598074  0.008598074  0.006113372  0.006113372 
#>          V19          V20          V26          V35          V47          V51 
#>  0.006113372  0.006113372  0.006113372  0.006113372  0.006113372  0.006113372 
#>          V55          V58          V59           V6          V31           V4 
#>  0.006113372  0.006113372  0.006113372  0.006113372  0.003896760  0.003896760 
#>          V14          V18           V2          V24          V25          V28 
#>  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000 
#>          V29           V3          V30          V38          V41          V42 
#>  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000 
#>          V45          V46          V48          V49           V5          V54 
#>  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000 
#>          V56          V57          V60           V7           V8          V13 
#>  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000 -0.002329015 
#>          V39          V40          V53          V15          V33          V43 
#> -0.002329015 -0.002329015 -0.002329015 -0.006067231 -0.006067231 -0.006067231 
#>          V23          V32          V44          V50          V22          V34 
#> -0.008507081 -0.008507081 -0.008507081 -0.012089351 -0.014638518 -0.014638518 

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

# Score the predictions
predictions$score()
#> classif.ce 
#> 0.08695652