CondInfTree <- ctree(DEFAULT ~ .,
data = train)
Using the above code, I get DEFAULT binary predictions which is what I'm seeking.
If I use:
CondInfTree <- ctree(Surv(DURATION,DEFAULT)~ .,
data = train)
I get time until default (Kaplan-Meier) predictions.
My goal is to include DURATION time as a variable but I want to predict whether a new, unseen, observation would default.
Would using the first tree work or would I be ignoring censored data that way?
... or is there something I'm missing altogether?
