R help - Jul 2005 - getting a variable from an object named using paste

Hello,

The subject could be articulated better ;(, but I am stuck...

In any case, my problem is the following: I am trying to use knn, and  
it requires a classification. So I am using cmeans for this. What I  
want to do is add the classification as a new column in my data.frame  
(to be used for knn). Now, the trick is that I would like to name the  
output of the cmeans classification based on the number of clusters  
(e.g. cmeans30 for 30 clusters). To do this, I do the following:


[The file for this example can be downloaded from:

http://mit.edu/costas/www/station-1-120103.txt   ]


library(e1071)
lala <- read.csv("station-1-120103.txt",header=F)
ll <- cbind(lala$V25,lala$V22)
num.of.clust=50
assign(paste("cmeans",num.of.clust,sep=""),cmeans(ll,num.of.clust))

and sure enough if I type cmeans50 I get the output of this cmeans run.

This also works, naturally:

llc<-cbind(ll,cmeans50$cluster)

[and I now get three columns, where the third column is the cluster ID.]

However, when I do what I really want (i.e. not call cmeans50  
explicitly, but through the name that I constructed programmatically):

llc <-
cbind(ll,(paste("cmeans",num.of.clust,sep=""))$cluster)

I only get two columns (the last term has no effect).

Any ideas as to what I am doing wrong?

Thanks a lot!

Costas


--
Constantinos Antoniou, Ph.D.
Department of Transportation Planning and Engineering
National Technical University of Athens
5, Iroon Polytechniou str. GR-15773, Athens, Greece

Hello,

The subject could be articulated better ;(, but I am stuck...

In any case, my problem is the following: I am trying to use knn, and  
it requires a classification. So I am using cmeans for this. What I  
want to do is add the classification as a new column in my data.frame  
(to be used for knn). Now, the trick is that I would like to name the  
output of the cmeans classification based on the number of clusters  
(e.g. cmeans30 for 30 clusters). To do this, I do the following:


[The file for this example can be downloaded from:

http://mit.edu/costas/www/station-1-120103.txt   ]


library(e1071)
lala <- read.csv("station-1-120103.txt",header=F)
ll <- cbind(lala$V25,lala$V22)
num.of.clust=50
assign(paste("cmeans",num.of.clust,sep=""),cmeans(ll,num.of.clust))

and sure enough if I type cmeans50 I get the output of this cmeans run.

This also works, naturally:

llc<-cbind(ll,cmeans50$cluster)

[and I now get three columns, where the third column is the cluster ID.]

However, when I do what I really want (i.e. not call cmeans50  
explicitly, but through the name that I constructed programmatically):

llc <-
cbind(ll,(paste("cmeans",num.of.clust,sep=""))$cluster)

I only get two columns (the last term has no effect).

Any ideas as to what I am doing wrong?

Thanks a lot!

Costas



-- 
Constantinos Antoniou, Ph.D.
Massachusetts Institute of Technology
Intelligent Transportation Systems Program
77 Massachusetts Ave., Rm. 1-249, Cambridge, MA 02139

Constantinos Antoniou wrote:
> Hello,
> 
> The subject could be articulated better ;(, but I am stuck...
> 
> In any case, my problem is the following: I am trying to use knn, and  
> it requires a classification. So I am using cmeans for this. What I  
> want to do is add the classification as a new column in my data.frame  
> (to be used for knn). Now, the trick is that I would like to name the  
> output of the cmeans classification based on the number of clusters  
> (e.g. cmeans30 for 30 clusters). To do this, I do the following:
> 
> 
> [The file for this example can be downloaded from:
> 
> http://mit.edu/costas/www/station-1-120103.txt   ]
> 
> 
> library(e1071)
> lala <- read.csv("station-1-120103.txt",header=F)
> ll <- cbind(lala$V25,lala$V22)
> num.of.clust=50
>
assign(paste("cmeans",num.of.clust,sep=""),cmeans(ll,num.of.clust))
> 
> and sure enough if I type cmeans50 I get the output of this cmeans run.
> 
> This also works, naturally:
> 
> llc<-cbind(ll,cmeans50$cluster)
> 
> [and I now get three columns, where the third column is the cluster ID.]
> 
> However, when I do what I really want (i.e. not call cmeans50  
> explicitly, but through the name that I constructed programmatically):
> 
> llc <-
cbind(ll,(paste("cmeans",num.of.clust,sep=""))$cluster)
This is a FAQ: use get()

Uwe Ligges

> I only get two columns (the last term has no effect).
> 
> Any ideas as to what I am doing wrong?
> 
> Thanks a lot!
> 
> Costas
> 
> 
>

Indeed. Thanks and sorry,

Costas

PS. also, sorry about sending the original post twice... I did not  
want to pollute the list any more with an apology then...


On 09 佄檨蝸縼蠀佄伝 2005, at 3:30 佄渷螠, Uwe Ligges wrote:
> Constantinos Antoniou wrote:
>
>> Hello,
>> The subject could be articulated better ;(, but I am stuck...
>> In any case, my problem is the following: I am trying to use knn,  
>> and  it requires a classification. So I am using cmeans for this.  
>> What I  want to do is add the classification as a new column in my  
>> data.frame  (to be used for knn). Now, the trick is that I would  
>> like to name the  output of the cmeans classification based on the  
>> number of clusters  (e.g. cmeans30 for 30 clusters). To do this, I  
>> do the following:
>> [The file for this example can be downloaded from:
>> http://mit.edu/costas/www/station-1-120103.txt   ]
>> library(e1071)
>> lala <- read.csv("station-1-120103.txt",header=F)
>> ll <- cbind(lala$V25,lala$V22)
>> num.of.clust=50
>>
assign(paste("cmeans",num.of.clust,sep=""),cmeans(ll,num.of.clust))
>> and sure enough if I type cmeans50 I get the output of this cmeans  
>> run.
>> This also works, naturally:
>> llc<-cbind(ll,cmeans50$cluster)
>> [and I now get three columns, where the third column is the  
>> cluster ID.]
>> However, when I do what I really want (i.e. not call cmeans50   
>> explicitly, but through the name that I constructed  
>> programmatically):
>> llc <-
cbind(ll,(paste("cmeans",num.of.clust,sep=""))$cluster)
>>
>
> This is a FAQ: use get()
>
> Uwe Ligges
>
>
>
>> I only get two columns (the last term has no effect).
>> Any ideas as to what I am doing wrong?
>> Thanks a lot!
>> Costas
>>
>
>
>


-- 
Constantinos Antoniou, Ph.D.
Massachusetts Institute of Technology
Intelligent Transportation Systems Program
77 Massachusetts Ave., Rm. 1-249, Cambridge, MA 02139