Re: [R] Nomogram (rms) for model with shrunk coefficients

2013-06-24 Thread David Winsemius
On Jun 24, 2013, at 12:00 PM, Sander van Kuijk wrote: > Dear R-users, > > I have used the nomogram function from the rms package for a logistic > regresison model made with lrm(). Everything works perfectly (r version > 2.15.1 on a mac). My question is this: if my final model is not the one > cr

Re: [R] Nomogram (rms) for model with shrunk coefficients

2013-06-24 Thread Frank Harrell
You are using an informal shrinkage method. It is much better to use penalized maximum likelihood estimation, built in to lrm. If you really want to go the informal route, compute the linear predictor from your final estimates and use ols( ) to predict that from the component variables (you'l

[R] Nomogram (rms) for model with shrunk coefficients

2013-06-24 Thread Sander van Kuijk
Dear R-users, I have used the nomogram function from the rms package for a logistic regresison model made with lrm(). Everything works perfectly (r version 2.15.1 on a mac). My question is this: if my final model is not the one created by lrm, but I internally validated the model and 'shrunk' the