GlenB
2010-Jan-18 10:11 UTC
[R] does any package have the functionality of ace() from old acepack?
I found the old ace function (from acepack) valuable a couple of years ago in helping to find a transformation of the response to approximate additivity in smooth functions of the predictors. ace used alternating conditional expectations, but I'm not overly fixated on algorithms as long as it works pretty well. Is there another package that has that sort of functionality? I haven't been able to locate one, but maybe I missed something. -- View this message in context: http://n4.nabble.com/does-any-package-have-the-functionality-of-ace-from-old-acepack-tp1016509p1016509.html Sent from the R help mailing list archive at Nabble.com.
Jonathan Baron
2010-Jan-18 11:15 UTC
[R] does any package have the functionality of ace() from old acepack?
Acepack is still in the archive section of CRAN: http://cran.r-project.org/src/contrib/Archive/acepack/acepack_1.3-2.2.tar.gz It seems to work despite the warning, but I guess you'd better test it. It is very simple: a few very old fortran scripts and an R wrapper by Thomas Lumley. I'm going to look into what needs to be done to bring it up to date, but no promises. A similar function is transace from the Hmisc package: http://finzi.psych.upenn.edu/library/Hmisc/html/transace.html I'm not sure that this does the same thing as ace. Jon On 01/18/10 02:11, GlenB wrote:> > > I found the old ace function (from acepack) valuable a couple of years ago > in helping to find a transformation of the response to approximate > additivity in smooth functions of the predictors. > ace used alternating conditional expectations, but I'm not overly fixated on > algorithms as long > as it works pretty well. > > Is there another package that has that sort of functionality? I haven't been > able to locate one, > but maybe I missed something.-- Jonathan Baron, Professor of Psychology, University of Pennsylvania Home page: http://www.sas.upenn.edu/~baron