There are several issues. First the Cox model is not for a binary outcome.
It is for a time-to-event outcome whose status (event vs. censored) is
binary. Second, split-sample validation does not work well with n < 2
in the combined sample. Third, reclassification tables are not used to
vali
As describe in the Hmisc package's improveProb function, that function is for
binary Y. And it's best to use category-free methods.
Frank
Petergodsk wrote
> Hi
>
> I'm new to R.
>
> Is it possible to use the improveProb function to generate categorybased
> NRI using a Cox model?
> I believe I
2 matches
Mail list logo
