Gradient Boosting with Regression Trees Survival Learner

Calls mboost::blackboost().

• lp is predicted by mboost::predict.mboost()

• distr is predicted by mboost::survFit() which assumes a PH fit with a Breslow estimator

• crank is identical to lp

The dist parameter is specified slightly differently than in mboost. Whereas the latter takes in objects, in this learner instead a string is specified in order to identify which distribution to use. As the default in mboost is the Gaussian family, which is not compatible with survival models, instead we have by default "coxph".

If the value given to the Family parameter is "custom.family" then an object of class mboost::Family() needs to be passed to the custom.family parameter.

Format

R6::R6Class() inheriting from LearnerSurv.

Construction

LearnerSurvBlackboost$new()
mlr_learners$get("surv.blackboost")
lrn("surv.blackboost")

Meta Information

• Type: "surv"

• Predict Types: distr, crank, lp

• Feature Types: integer, numeric, factor

• Packages: mboost distr6 survival partykit mvtnorm

References

Peter Buehlmann and Torsten Hothorn (2007), Boosting algorithms: regularization, prediction and model fitting. Statistical Science, 22(4), 477–505.

Torsten Hothorn, Kurt Hornik and Achim Zeileis (2006). Unbiased recursive partitioning: A conditional inference framework. Journal of Computational and Graphical Statistics, 15(3), 651–674.

Yoav Freund and Robert E. Schapire (1996), Experiments with a new boosting algorithm. In Machine Learning: Proc. Thirteenth International Conference, 148–156.

Jerome H. Friedman (2001), Greedy function approximation: A gradient boosting machine. The Annals of Statistics, 29, 1189–1232.

Greg Ridgeway (1999), The state of boosting. Computing Science and Statistics, 31, 172–181.

See also

Other survival learners: LearnerSurvCVGlmnet, LearnerSurvCoxPH, LearnerSurvFlexible, LearnerSurvGBM, LearnerSurvGamboost, LearnerSurvGlmboost, LearnerSurvGlmnet, LearnerSurvKaplan, LearnerSurvMboost, LearnerSurvNelson, LearnerSurvParametric, LearnerSurvPenalized, LearnerSurvRandomForestSRC, LearnerSurvRanger, LearnerSurvRpart, LearnerSurvSVM

Examples

library(mlr3)
task = tgen("simsurv")$generate(200)
learner = lrn("surv.blackboost")
resampling = rsmp("cv", folds = 3)
resample(task, learner, resampling)
#> <ResampleResult> of 3 iterations
#> * Task: simsurv
#> * Learner: surv.blackboost
#> * Warnings: 0 in 0 iterations
#> * Errors: 0 in 0 iterations