[R] Fitting weibull and exponential distributions to left censoring data

Terry Therneau therneau at mayo.edu
Fri Oct 31 14:27:03 CET 2008


 Use the survreg function.
 
 There are many different ways to parameterize a Weibull.  The survreg function 
imbeds it a general location-scale familiy, which is a different 
parameterization than the rweibull function. 
 
> y <- rweibull(1000, shape=2, scale=5)
> survreg(Surv(y)~1, dist="weibull")

Coefficients:
(Intercept) 
   1.592543 

Scale= 0.5096278 

Loglik(model)= -2201.9   Loglik(intercept only)= -2201.9

----

 survreg's scale  =    1/(rweibull shape)
 survreg's intercept = log(rweibull scale)
 For the log-likelihood all parameterizations lead to the same value.

  There is not "right" or "wrong" parameterization for a Weibull (IMHO), but 
there certainly is a lot of room for confusion.  This comes up enough that I 
have just added it as an example in the survreg help page, which will migrate to 
the general R distribution in due course.
  
  Terry Therneau



More information about the R-help mailing list