As described in official documentation at the part of basehaz.gbm:
Arguments:
cumulative, If TRUE the cumulative survival function will be computed.
...
Details:
The proportional hazard model assumes h(t|x)=lambda(t)*exp(f(x)). gbm can estimate the f(x) component via partial likelihood. After estimating f(x), basehaz.gbm can compute the a nonparametric estimate.
Is "the cumulative survival function will be computed" correct? Is there terminology error ? Since the concept of cumulative survival function and cumulative hazard function is different.
I read the code implemented in gbm-baseline-hazard.r and run an example for this function with setting t.eval=c(10, 20, 30), but I got three values in ascending order, which is not consistent to the properties of survival function, so I opened this issue, and same doubts has also been proposed at stackoverflow
If this issue is actually existing, I think it would be a wrong guide to someone like me who read the documentation, and the correct version should be "cumulative, If TRUE the cumulative baseline hazard function will be computed."
As described in official documentation at the part of basehaz.gbm:
Is "the cumulative survival function will be computed" correct? Is there terminology error ? Since the concept of cumulative survival function and cumulative hazard function is different.
I read the code implemented in
gbm-baseline-hazard.rand run an example for this function with settingt.eval=c(10, 20, 30), but I got three values in ascending order, which is not consistent to the properties of survival function, so I opened this issue, and same doubts has also been proposed at stackoverflowIf this issue is actually existing, I think it would be a wrong guide to someone like me who read the documentation, and the correct version should be "cumulative, If TRUE the cumulative baseline hazard function will be computed."
thx!