Calculating Estimated Population Survival Curve

[puckybreg]

Hi Pycox Community!

I am working on a project applying survival analysis to the dwell times of insects and have found the PyCox package quite useful.
I have mostly been following along with the simplest example from 01_introduction.ipynb but have had to make some inferences on the Population Survival Curve.

In the tutorial, the predicted survival curves of the test data are built and the graphed indivudally. I am hoping to come up with an estimated Population Survival Curve. Is this as simple as taking the mean of our predicted survival curves?

surv = model.predict_surv_df(x_test)
surv['mean'] = surv.mean(axis=1)
surv['mean'].plot(drawstyle='steps-post')

Let me know your thoughts.

[sourasen1011]

Not sure if this is the answer you are looking for, but if you are looking for a Population survival curve, sounds like a non-parametric estimator such as the Kaplan Meier or Nelson Aalen may be good for you. The Kaplan Meier one doesn't consider any covariates - it just plots the survival at each observed event time as the cumulative product of previous survivals (starting with 1, of course).

[alexxiecan]

Hello, deer puckybreg and everyone
The demo of the deepsurvey model draws an individual survival curve, and I also hope to be able to draw a curve for the entire population. Do you know if anyone else has a good solution?
sincerely !!!! thanks

[alexxiecan]

dear

[shahidhaider-altis]

I concur with @sourasen1011
You are looking for something like a KM curve, which is a non-parametric estimator for the population survival curve