11.7.5 Survival Trees

In Chapter 8, we discussed flexible and adaptive learning procedures such as trees, random forests, and boosting, which we applied in both the regression and classification settings. Most of these approaches can be generalized to theof classification and regression trees that use a split criterion that maximizessurvival analysis setting. For example, survival trees are a modification survivaltrees

trees

11.8 Lab: Survival Analysis 489

the difference between the survival curves in the resulting daughter nodes. Survival trees can then be used to create random survival forests.