[R] Shared Frailty in survival package (left truncation, time-dep. covariates)

[Stefan Pohl]

Dear list,

I want o fit a shared gamma frailty model with the frailty specification in the 
survival package.

I have partly left-truncated data and time-dependent covariates. Is it possible 
to
combine these two things in the frailty function. Or are the results wrong if I 
use data in the start-stop-formulation which account for delayed entry?

Is the frailty distribution updated in the left-truncated case?

Or is the frailty function only built for time-constant covariates and left 
truncated data?

Thank you, 

Stef.

[[alternative HTML version deleted]]

__
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html

[R] nlminb to optmin

[Stefan Pohl]

Hi!

I want to convert S-Plus 6.2 code to R 2.1.0. Instead of the function nlminb I 
use the function optmin

optmin(start,fn,gr,method=L-BFGS-B, lower, upper, hess,...)

But then I get the Error in optmin ...: L-BFGS-B needs finite values of fn

Then I used optmin(start,fn,gr,method=BFGS, hess, ...)

But then I get the Error in optmin ...: initial value in vmmin is not finite

I know the final parameter estimates from S-Plus which I use as starting values 
in R.
The upper and lower bounds are close around the final estimates.
So there is not much to maximize.

What can I do?

Thank you for help,

Peter

[[alternative HTML version deleted]]

__
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html

[R] Left truncation in shared frailty models with time-varying covariates

[Stefan Pohl]

Hi!

I want to estimate a shared gamma frailty model with left truncated data. I use 
a parametric baseline hazard so that I can use
simple ML estimation. As I have a big data set it is ok to assume piecewise 
constant baseline hazards. 

As my data are left truncated I have modified the definition of the risk set.

Do I also have to modifiy the frailty distribution if I have left truncated 
data? And if I have to, how? And if I have to, are time-varying covariates 
modeled with the method of episode splitting a problem?

Thank you for any comments,

Stefan.



[[alternative HTML version deleted]]

__
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html

[R] transform normally distributed random terms to gamma distributed random terms

[Stefan Pohl]

Hi,

I have normally distributed random terms u~N(0,1). I want to get gamma 
distributed random terms g~(scale,shape) with
E(g)=1=shape/scale and var(g)=theta=1/scale=1/shape. 

How can I reach my goal? The following way doesn't work: use the distribution 
function of u to get U(0,1)- distributed random
terms, then take the quantile function of the gamma distribution with shape and 
scale.

The resulting random terms must be ~gamma(shape, scale).

But it doesn't work.

Is there a mistake or do you know another way?

Thanks, Stefan.
[[alternative HTML version deleted]]

__
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html