R help - Aug 2005 - Penalized likelihood-ratio chi-squared statistic: L.R. model for Goodness of fit?

Dear R list,

 

From the lrm() binary logistic model we derived the G2 value or the
likelihood-ratio chi-squared statistic given as L.R. model, in the output of
the lrm().

 

 

How can this value be penalized for non-linearity (we used splines in the
lrm function)?

 

lrm.iRVI <- lrm(arson ~ rcs(iRVI,5),
penalty=list(simple=10,nonlinear=100,nonlinear.interaction=4)) 

 

This didn’t work properly.

 

 

The aim is to obtain a value that can be used to compare the goodness of fit
of the different univariate binary logistic models. 

 

(The lower the value, the better the fit)

 

 

Kind regards,

Jan

____________________________________________________________________
Ir. Jan Verbesselt
Research Associate
Group of Geomatics Engineering
Department Biosystems ~ M³-BIORES
Vital Decosterstraat 102, 3000 Leuven, Belgium
Tel: +32-16-329750   Fax: +32-16-329760
http://gloveg.kuleuven.ac.be/
_______________________________________________________________________

 


	[[alternative HTML version deleted]]

Jan Verbesselt wrote:
> Dear R list,
> 
>  
> 
> From the lrm() binary logistic model we derived the G2 value or the
> likelihood-ratio chi-squared statistic given as L.R. model, in the output
of
> the lrm().
> How can this value be penalized for non-linearity (we used splines in the
> lrm function)?
> 
> lrm.iRVI <- lrm(arson ~ rcs(iRVI,5),
> penalty=list(simple=10,nonlinear=100,nonlinear.interaction=4)) 
> 
> This didn抰 work properly.
Please following the posting guide.  What do you mean by 'work' and what
is the output?

You are attempting to penalize for nonexistent interaction terms.

Differential penalization is only appropriate if there are many similar 
terms being penalized (e.g., you fit a multivariable model and want to 
penalize all nonlinear terms in all variables combined).
> The aim is to obtain a value that can be used to compare the goodness of
fit
> of the different univariate binary logistic models. 
By univariate I assume you mean univariable.  Penalization is primarily 
used to fit multivariable models.  It allows you to fit "bigger"
models.

But for your purpose comparing AIC of various models might be entertained.

Frank
> 
> (The lower the value, the better the fit)
> 
> Kind regards,
> 
> Jan
> 
> ____________________________________________________________________
> Ir. Jan Verbesselt
> Research Associate
> Group of Geomatics Engineering
> Department Biosystems ~ M伋-BIORES
> Vital Decosterstraat 102, 3000 Leuven, Belgium
> Tel: +32-16-329750   Fax: +32-16-329760
> http://gloveg.kuleuven.ac.be/
> _______________________________________________________________________
> 
>  
> 
> 
> 	[[alternative HTML version deleted]]
> 
> 
> 
> ------------------------------------------------------------------------
> 
> ______________________________________________
> R-help at 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

-- 
Frank E Harrell Jr   Professor and Chair           School of Medicine
                      Department of Biostatistics   Vanderbilt University