[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: Compare predicted survival function with the observed
I have estimated a piece-wise constant exponential model using streg. My doubts are: 1) How could I test the predictive power of the model? Could I compare the predicted survival function with the observed (Kaplan-Meier estimate)? I have tried to use the Stata macro stexpect downloadable from SSC but I do not undertang what "ratevar" means. I do not have a threshold probability that I should use, so I think that I could get it using sts gen km=s. Is that correct? 2) After using stexpect and getting the output, what I have to do?