Skip to content
Prev 13320 / 398502 Next

Fitting distributions

On Wed, 5 Sep 2001, Cowpertwait, Paul wrote:

            
See survreg in package survival. It's a trivial case (no regressors, no
censoring).

survreg(y ~ 1, dist="weibull")

survreg(log(y) ~ 1, dist="Weibull")

since log-Weibulls are Gumbel.

Direct computation of the likelihood is easy: see the recent postings here,
or V&R's exercises (e.g 8.10).
I think you will need to program that yourself.