R help - Feb 2011 - survfit - help avoiding re-inventing the wheel.

Greetings, R-ians:

 

I would like to calculate the loglikelihood based on a Surv object and
non-mle parameter estimates.  I already have the mles as provided by
survreg( Surv(time, time2, event, type) ).

 

I can compute a loglikelihood based on a given density, censoring types, and
parameter values, but find myself doing what must already be available, if I
knew how to access it.

 

The reason for wanting to compute a loglikelihood at non-mle values is to
estimate the log likelihood ratios.

 

Any help is much appreciated.

 

Thanks.

 

Charles Annis, P.E.

 

Charles.Annis@StatisticalEngineering.com

561-352-9699

http://www.StatisticalEngineering.com

 


	[[alternative HTML version deleted]]