Gradient Boosting with Regression Trees Survival Learner

Gradient boosting with regression trees for survival analysis. Calls mboost::blackboost() from mboost.

Details

distr prediction made by mboost::survFit().

Prediction types

This learner returns two to three prediction types:

1. lp: a vector containing the linear predictors (relative risk scores), where each score corresponds to a specific test observation. Calculated using mboost::predict.blackboost(). If the family parameter is not "coxph", -lp is returned, since non-coxph families represent AFT-style distributions where lower lp values indicate higher risk.

2. crank: same as lp.

3. distr: a survival matrix in two dimensions, where observations are represented in rows and time points in columns. Calculated using mboost::survFit(). This prediction type is present only when the family distribution parameter is equal to "coxph" (default). By default the Breslow estimator is used for computing the baseline hazard.

Dictionary

This Learner can be instantiated via lrn():

lrn("surv.blackboost")

Meta Information

• Task type: “surv”

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

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

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

Parameters

Id	Type	Default	Levels	Range
family	character	coxph	coxph, weibull, loglog, lognormal, gehan, cindex, custom	-
custom.family	untyped	-		-
nuirange	untyped	c(0, 100)		-
offset	untyped	-		-
center	logical	TRUE	TRUE, FALSE	-
mstop	integer	100		\([0, \infty)\)
nu	numeric	0.1		\([0, 1]\)
risk	character	-	inbag, oobag, none	-
stopintern	logical	FALSE	TRUE, FALSE	-
trace	logical	FALSE	TRUE, FALSE	-
oobweights	untyped	-		-
teststat	character	quadratic	quadratic, maximum	-
splitstat	character	quadratic	quadratic, maximum	-
splittest	logical	FALSE	TRUE, FALSE	-
testtype	character	Bonferroni	Bonferroni, MonteCarlo, Univariate, Teststatistic	-
maxpts	integer	25000		\([1, \infty)\)
abseps	numeric	0.001		\((-\infty, \infty)\)
releps	numeric	0		\((-\infty, \infty)\)
nmax	untyped	-		-
alpha	numeric	0.05		\([0, 1]\)
mincriterion	numeric	0.95		\([0, 1]\)
logmincriterion	numeric	-0.05129329		\((-\infty, 0]\)
minsplit	integer	20		\([0, \infty)\)
minbucket	integer	7		\([0, \infty)\)
minprob	numeric	0.01		\([0, 1]\)
stump	logical	FALSE	TRUE, FALSE	-
lookahead	logical	FALSE	TRUE, FALSE	-
MIA	logical	FALSE	TRUE, FALSE	-
nresample	integer	9999		\([1, \infty)\)
tol	numeric	1.490116e-08		\([0, \infty)\)
maxsurrogate	integer	0		\([0, \infty)\)
mtry	integer	-		\([0, \infty)\)
maxdepth	integer	-		\([0, \infty)\)
multiway	logical	FALSE	TRUE, FALSE	-
splittry	integer	2		\([1, \infty)\)
intersplit	logical	FALSE	TRUE, FALSE	-
majority	logical	FALSE	TRUE, FALSE	-
caseweights	logical	TRUE	TRUE, FALSE	-
sigma	numeric	0.1		\([0, 1]\)
ipcw	untyped	1		-
na.action	untyped	stats::na.omit		-

References

Bühlmann, Peter, Yu, Bin (2003). “Boosting with the L 2 loss: regression and classification.” Journal of the American Statistical Association, 98(462), 324–339.

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 -> LearnerSurvBlackBoost

Methods

Public methods

• LearnerSurvBlackBoost$new()

• LearnerSurvBlackBoost$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

LearnerSurvBlackBoost$new()

---

Method clone()

The objects of this class are cloneable with this method.

Usage

LearnerSurvBlackBoost$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# Define the Learner
learner = mlr3::lrn("surv.blackboost")
print(learner)
#> <LearnerSurvBlackBoost:surv.blackboost>: Gradient Boosting
#> * Model: -
#> * Parameters: family=coxph
#> * Packages: mlr3, mlr3proba, mlr3extralearners, mboost, pracma
#> * Predict Types:  [crank], distr, lp
#> * Feature Types: integer, numeric, factor
#> * Properties: 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$model)
#> 
#> 	 Model-based Boosting
#> 
#> Call:
#> mboost::blackboost(formula = task$formula(task$feature_names),     data = task$data(), family = family)
#> 
#> 
#> 	 Cox Partial Likelihood 
#> 
#> Loss function:  
#> 
#> Number of boosting iterations: mstop = 100 
#> Step size:  0.1 
#> Offset:  0 
#> Number of baselearners:  1 
#> 


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

# Score the predictions
predictions$score()
#> surv.cindex 
#>   0.8550345