[R] parametric proportional hazard regression
Valentin Dimitrov
vsdimitrov at yahoo.com
Sat Jul 8 14:24:05 CEST 2006
--- Frank E Harrell Jr <f.harrell at vanderbilt.edu>
wrote:
> That will be a very strange model that I've never
> seen used before in
> survival analysis. Interpretation of parameters
> other than the hazard
> ratio may be tricky. Why do you need a
> nontraditional model such as this?
>
> Frank
>
> --
> Frank E Harrell Jr Professor and Chair
> School of Medicine
> Department of Biostatistics
> Vanderbilt University
>
I am new to the field of survival analysis and my
first text read on that topic is Franses/Paap,
Quantitative Models in Marketing Research, Chapter 8.
There the kind of models I describe are the standard
proportional hazard models. But also in Internet I
find texts which say something about parametric PH
models, where the baseline hazard can take some
specific form! See for instance:
http://www.weibull.com/AccelTestWeb/proportional_hazards_model.htm
The second part reads "parametric PH models" and there
is an example with the Weibull baseline distribution.
The log-likelihood is derived, too. The same could be
done with the other distributions I mentioned
(lognormal, loglogistic, exponential).
Anyway, if you say these models are untraditional,
then maybe there is also no R-function for them...
In that case I'll write a R-code myself to fit them
using ML.
Kind regards,
Valentin
More information about the R-help
mailing list