Skip to content

Logistic Regression on a Windows Machine

3 messages · Carrie Beam, Deepayan Sarkar, Brian Ripley

Original
Carrie Beam
# 
Hello everybody,

I am hoping you can help me.  I just downloaded and installed the 
"precompiled binary distribution" of R to my Windows machine (running 
Windows Millennium).  (I did not download or install anything else - no 
packages).

I need to run a logistic regression and understand the function I need is 
polr() or maybe SSlogis().

However, when I try to invoke either function, I get
    Error: couldn't find function "polr"
    Error: couldn't find function "SSlogis"

I have done some snooping on the website and think I may need the MASS 
package for polr() and the NLS package for SSlogis().

My questions:
1.  Are these indeed the functions I need to run a logistic regression?
2.  If so, are the packages with them available for a Windows machine?

Thanks a million in advance -- you all are the greatest
-- Carrie
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Deepayan Sarkar
#
On Sunday 24 November 2002 08:51 pm, Carrie Beam wrote:
Not sure about your first question, depends on what you want to do (standard 
logistic regression can be done with the glm function, while polr does 
proportional odds logistic regression).

As for R packages (more commonly referred to as libraries), many of them come 
with the default R installation, but you have to load them before you can use 
them. MASS and nls are among them. You should see a full list using

library()

For more information, try

?glm
library(MASS)
?polr

etc.

Deepayan


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Brian Ripley
#
On Sun, 24 Nov 2002, Carrie Beam wrote:

            

      
      
      
    

        
Probably not.  See ?glm if you mean a binomial logistic regression.
nls/SSlogis will give you a least-squares logistic fit.

            

      
      
      
    

        
Yes, just look at the results of library().