Cross validation on fitted survival objects?

Question

I can see how cv.glm work with a glm object, but what about fitted survival models?

I have a bunch of models (Weibull, Gompertz, lognormal, etc). I want to assess the prediction error using cross validation. Which package/function can do this in R?

Answer 1

SuperLearner can do V-fold cross-validation for a large library of underlying machine learning algorithms, not sure that it includes survival models. Alternatively, take a look at the cvTools package, which is designed to help do cross-validation of any prediction algorithm you give it.