Survival L1 and L2 Penalized Cox Learner

Penalized (L1 and L2) Cox Proportional Hazards model. Calls penalized::penalized() from penalized.

Details

The penalized and unpenalized arguments in the learner are implemented slightly differently than in penalized::penalized(). Here, there is no parameter for penalized but instead it is assumed that every variable is penalized unless stated in the unpenalized parameter.

Prediction types

This learner returns two prediction types:

1. distr: a survival matrix in two dimensions, where observations are represented in rows and time points in columns. Calculated using the internal penalized::predict() function. By default the Breslow estimator penalized::breslow() is used for computing the baseline hazard.

2. crank: the expected mortality using mlr3proba::.surv_return().

Initial parameter values

• trace is set to "FALSE" to disable printing output during model training.

Dictionary

This Learner can be instantiated via lrn():

lrn("surv.penalized")

Meta Information

• Task type: “surv”

• Predict Types: “crank”, “distr”

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

• Required Packages: mlr3, mlr3proba, mlr3extralearners, penalized, pracma

Parameters

Id	Type	Default	Levels	Range
unpenalized	untyped	-		-
lambda1	untyped	0		-
lambda2	untyped	0		-
positive	logical	FALSE	TRUE, FALSE	-
fusedl	logical	FALSE	TRUE, FALSE	-
startbeta	numeric	-		\((-\infty, \infty)\)
startgamma	numeric	-		\((-\infty, \infty)\)
steps	integer	1		\([1, \infty)\)
epsilon	numeric	1e-10		\([0, 1]\)
maxiter	integer	-		\([1, \infty)\)
standardize	logical	FALSE	TRUE, FALSE	-
trace	logical	TRUE	TRUE, FALSE	-

References

Goeman, J J (2010). “L1 penalized estimation in the Cox proportional hazards model.” Biometrical journal, 52(1), 70–84.

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/basics.html#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

RaphaelS1

Super classes

mlr3::Learner -> mlr3proba::LearnerSurv -> LearnerSurvPenalized

Methods

Public methods

• LearnerSurvPenalized$new()

• LearnerSurvPenalized$selected_features()

• LearnerSurvPenalized$clone()

Inherited methods

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

---

Method new()

Creates a new instance of this R6 class.

Usage

LearnerSurvPenalized$new()

---

Method selected_features()

Selected features are extracted with the method coef() of the S4 model object, see penalized::penfit(). By default it returns features with non-zero coefficients.

Note: Selected features can be retrieved only for datasets with numeric features, as the presence of factors with multiple levels makes it difficult to get the original feature names.

Usage

LearnerSurvPenalized$selected_features()

Returns

character().

---

Method clone()

The objects of this class are cloneable with this method.

Usage

LearnerSurvPenalized$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# Define the Learner
learner = mlr3::lrn("surv.penalized")
print(learner)
#> <LearnerSurvPenalized:surv.penalized>: Penalized Regression
#> * Model: -
#> * Parameters: trace=FALSE
#> * Packages: mlr3, mlr3proba, mlr3extralearners, penalized, pracma
#> * Predict Types:  [crank], distr
#> * Feature Types: logical, integer, numeric, factor
#> * Properties: -

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

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

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

print(learner$model)
#> $model
#> Penalized cox regression object
#> 6 regression coefficients
#> 
#> Loglikelihood =	 -1319.822 
#> 
#> $task_has_factors
#> [1] FALSE
#> 


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

# Score the predictions
predictions$score()
#> surv.cindex 
#>   0.8361721