Let me add to what Thomas said.
The default way of getting the coefficents from any function returning a
more complex statistical model is coef.
from win32com.client import Dispatch
r=Dispatch("StatConnectorSrv.StatConnector")
r.Init("R")
r.EvaluateNoReturn("data(iris)")
r.EvaluateNoReturn("iris.lm <- lm(Sepal.Length ~ Sepal.Width,
data = iris)")
res=Evaluate("coef(iris.lm)")
should assign the vector of coefficients to your
Thomas Baier wrote:
> David,
>
>> I am using Python 2.5 and R2.5.1
>>
>> I am trying to do a simple test using the Iris data set and a regression.
>>
>> Here is my code
>>
>>
>> from win32com.client import Dispatch
>> r=Dispatch("StatConnectorSrv.StatConnector")
>> r.Init("R")
>> r.Evaluate("data(iris)")
>> r.Evaluate("iris.lm <- lm(Sepal.Length ~ Sepal.Width, data = iris)")
>>
>>
>> If I do the above in R command line it works fine. In Python I get nasty
>> little errors and don't know why. WHen I do a print
>> print "Objects= ", r.Evaluate("objects()") command I see the iris data set
>> is loaded.....the regression just does not work...but as mentioned...same
>> code works in R command line.
>>
>> What is going on or am I missing some additional initializations?
>
> lm() returns an object of class "lm", which cannot be transferred
> directly using Evaluate(). Use EvaluateNoReturn() instead of Evaluate()
> and use Evaluate() to retrieve the transferrable parts of "iris.lm",
> e.g. r.Evaluate("iris.lm$coefficients")
>
> Thomas
> _______________________________________________
> Rcom-l mailing list
> Rcom-l@mailman.csd.univie.ac.at
> http://mailman.csd.univie.ac.at/mailman/listinfo/rcom-l
> More information (including a Wiki) at http://rcom.univie.ac.at
>
>
_______________________________________________
Rcom-l mailing list
Rcom-l@mailman.csd.univie.ac.at
http://mailman.csd.univie.ac.at/mailman/listinfo/rcom-l
More information (including a Wiki) at http://rcom.univie.ac.at
