Survival Cox Model with Likelihood Based Boosting Learner

Fit a Survival Cox model with a likelihood based boosting algorithm. Calls CoxBoost::CoxBoost() from package 'CoxBoost'.

Details

Use LearnerSurvCoxboost and LearnerSurvCVCoxboost for Cox boosting without and with internal cross-validation of boosting step number, respectively. Tuning using the internal optimizer in LearnerSurvCVCoxboost may be more efficient when tuning stepno only. However, for tuning multiple hyperparameters, mlr3tuning and LearnerSurvCoxboost will likely give better results.

Prediction types

This learner returns three prediction types, using the internal predict.CoxBoost() function:

1. lp: a vector containing the linear predictors (relative risk scores), where each score corresponds to a specific test observation.

2. crank: same as lp.

3. distr: a 2d survival matrix, with observations as rows and time points as columns. The internal transformation uses the Breslow estimator to compute the baseline hazard and compose the survival distributions from the lp predictions.

Dictionary

This Learner can be instantiated via lrn():

lrn("surv.coxboost")

Meta Information

• Task type: “surv”

• Predict Types: “crank”, “distr”, “lp”

• Feature Types: “integer”, “numeric”

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

Parameters

Id	Type	Default	Levels	Range
unpen.index	untyped	-		-
standardize	logical	TRUE	TRUE, FALSE	-
stepno	integer	100		\([0, \infty)\)
penalty	numeric	-		\((-\infty, \infty)\)
criterion	character	pscore	pscore, score, hpscore, hscore	-
stepsize.factor	numeric	1		\((-\infty, \infty)\)
sf.scheme	character	sigmoid	sigmoid, linear	-
pendistmat	untyped	-		-
connected.index	untyped	-		-
x.is.01	logical	FALSE	TRUE, FALSE	-
return.score	logical	TRUE	TRUE, FALSE	-
trace	logical	FALSE	TRUE, FALSE	-
at.step	untyped	-		-

Installation

The package 'CoxBoost' is not on CRAN and has to be installed from GitHub using remotes::install_github("binderh/CoxBoost").

References

Binder, Harald, Allignol, Arthur, Schumacher, Martin, Beyersmann, Jan (2009). “Boosting for high-dimensional time-to-event data with competing risks.” Bioinformatics, 25(7), 890–896.

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 -> LearnerSurvCoxboost

Methods

Public methods

• LearnerSurvCoxboost$new()

• LearnerSurvCoxboost$selected_features()

• LearnerSurvCoxboost$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

LearnerSurvCoxboost$new()

---

Method selected_features()

Returns the set of selected features which have non-zero coefficients. Calls the internal coef.CoxBoost() function.

Usage

LearnerSurvCoxboost$selected_features(at_step = NULL)

Arguments

at_step

(integer(1))
Which boosting step to get the coefficients for. If no step is given (default), the final boosting step is used.

Returns

(character()) vector of feature names.

---

Method clone()

The objects of this class are cloneable with this method.

Usage

LearnerSurvCoxboost$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# Define the Learner
learner = mlr3::lrn("surv.coxboost")
print(learner)
#> <LearnerSurvCoxboost:surv.coxboost>: Likelihood-based Boosting
#> * Model: -
#> * Parameters: list()
#> * Packages: mlr3, mlr3proba, mlr3extralearners, CoxBoost, pracma
#> * Predict Types:  [crank], distr, lp
#> * Feature Types: integer, numeric
#> * Properties: selected_features, weights

# 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)
#> 100 boosting steps resulting in 4 non-zero coefficients  
#> partial log-likelihood: -1103.542 


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

# Score the predictions
predictions$score()
#> surv.cindex 
#>   0.8290042