Skip to content

Generalized Additive Models: How to create publication-ready regression tables

5 messages · Emilio López, davidyeager, Frank E Harrell Jr

#
Hi - 

I have a series of 9 GAM regressions with about 5 parametric effects and
three non-parametric effects in each.  

What is a good library or command for turning GAM outputs into
publication-ready regression tables?

I tried apsrtable and the mtable command in memisc but neither seemed to
work with the gam output.

I'd be okay with two separate tables - one for the parametric effects and
one for the non-parametric effects.  

Best,

David

--
View this message in context: http://r.789695.n4.nabble.com/Generalized-Additive-Models-How-to-create-publication-ready-regression-tables-tp3884432p3884432.html
Sent from the R help mailing list archive at Nabble.com.
#
Thanks!

Yes, that produces tables that are formatted in the same way as the gam
output.  I'm hoping to have publication-ready tables that have standard
errors in parentheses listed below coefficients.  Do you know of a method to
do that?

David

--
View this message in context: http://r.789695.n4.nabble.com/Generalized-Additive-Models-How-to-create-publication-ready-regression-tables-tp3884432p3885238.html
Sent from the R help mailing list archive at Nabble.com.
#
In many cases a flexible parametric fit, using regression splines, will
result in a fit that is as good as a gam, with similar regression shapes. 
The rms package has a latex method that will represent such fits in
interpretable algebraic form.  latex(fit) does that, and print(fit,
latex=TRUE) will give a nice table though not formatted in the way you
described.
Frank
Emilio L?pez wrote:
-----
Frank Harrell
Department of Biostatistics, Vanderbilt University
--
View this message in context: http://r.789695.n4.nabble.com/Generalized-Additive-Models-How-to-create-publication-ready-regression-tables-tp3884432p3887519.html
Sent from the R help mailing list archive at Nabble.com.