Conditional Survival Curve :: Forecasting
Bminer
This is my first post and please excuse if I get a couple *easy* ones
out of the way in the next couple days...;)
I am looking at using survival curves to forecast the number of
individuals remaining at some point in time. This is for a business
purpose to estimate costs. I am building the survival models from time
0. However, when implementing the forecast we will be "scoring" the
existing database and there will be of course individuals which are
currently alive and have a certain tenure.
You cant say on day X that 100% of the cohort survives (because they
have), but at day X+1 that 60% survive (for example, if the
unconditional survival curve which was built from time 0 estimates
survival at day X as 0.60).
It seems I have seen a simple method to deal with this issue - I think
called conditional survival... Does anyone know how to best take into
account that a cohort has an existing tenure?
Thanks!
Brian