R help - Nov 2009 - multiple column argument in formula (MASS:lda)

Dear R Help team,

I'm having some problems running a discriminant analysis (MASS:lda)
over my data, i have a grouping variable coming from a previous
cluster analysis and several discriminators (78 actually). I'm trying
to run lda using column index as independent variable in the formula
like it's done in the help page for ?formula:
>lda(Grupo~paste(names(cien[,3:80]),collapse="+"),data=cien)
but i get the following error:

Error in model.frame.default(formula = Grupo ~ paste(names(cien[, 3:80]),  :
  different variable length (found in 'paste(names(cien[, 3:80]),
collapse = "+")')
(This is my english translation of the error msg while i use R in
spanish so it should look different in your machines)

The funny thing is that if i use:
>paste(names(cien[,3:80]),collapse="+")
[1] "X.SO.M2+X.SO.M4+X.SO.M6+X.SO.M8+X.SO.M10+X.SO.M12+.....

and then copy and paste the result in the formula:
> lda(Grupo~X.SO.M2+X.SO.M4+X.SO.M6+X.SO.M8+X.SO.M10+X.SO.M12+....,data=cien)
then there's no error!!! the lda runs just fine.

I know i might be doing something wrong but i don't know where so
could you please enlighten me? Thanks in advance

PD. Here's a piece of the data:

            NOMEST                      Grupo  X.SO.M2  X.SO.M4
X.SO.M6  X.SO.M8  .....(78 var)
1          Ci?naga La Ahuyama       3       95.1              152.3
115.3    97.6
2          Boca del Ca?o Grande     3       91.0               104.2
131.7    89.4
3          Buenavista                      3       115.6
102.9   122.0   110.5
4          Ci?naga El Loro               1       104.0
80.9    68.5    50.7
.
.
.


--
Marcos Antonio Carvajalino Fern?ndez
Estudiante de Ingenier?a Ambiental y Sanitaria
Universidad del Magdalena
Santa Marta, Colombia

Hello Marcos,
When you use paste(.), it will return a String, and the first argument of
lda(.) does not read a string as formula. 
>lda(Grupo~paste(names(cien[,3:80]),collapse="+"),data=cien)
Here's what I'd do: 

colIndex <- c(1,2,4,5,...) #what ever index of your input data set that you
want to use as variables

#I presume Grupo is the class column (col=3)?
lda(cien[,colIndex], cien[,3])

Hope this helps.. 

Wai-Kuan Yip
http://www.deakin.edu.au/~waiyip http://www.deakin.edu.au/~waiyip 


-- 
View this message in context:
http://n4.nabble.com/multiple-column-argument-in-formula-MASS-lda-tp931285p931814.html
Sent from the R help mailing list archive at Nabble.com.