Not really an answer but an idea that you could do is to vary the inputs randomly and measure the sorts of values generated. See how varied they are.
• 0 views
•
link
Hi, I am using R package 'randomForestSRC' to do survival analysis. I used rfsrc function to build the model with the training dataset. And I used predict function to get the predicted survival of the test set. It seems I can get the predicted days of survival. But is there a way to get some confidence level for each of these predictions as well?
Thanks!
No answers yet.
Log in to answer this question.