Skip to content

Silhouette function problem

5 messages · David Winsemius, jim holtman, ADias

ADias
#
ADias wrote:
Hi,

I have made this script

library(cluster)
d<-hclust(dist(iris[,-5]))
avgs<-sapply(1:20,function(x)
summary(silhouette(cutree(d,x),
dist(iris[,-5]))))
str(avgs)

I need $ avg.width  but I don't know how to take that out from the object
avgs.
The mode from avgs is "list" but it does not work as a normal list.

Any help would be apreciated.

thanks
A. Dias
David Winsemius
#
On Dec 31, 2010, at 10:34 AM, ADias wrote:

            

      
      
      
    

        
Do these examples of accessing <something> resembling what you are  
asking in that object help?

 > avgs[[2]]$clus.avg.widths
         1         2
0.3861393 0.6566469
 > avgs[[7]]$clus.avg.widths
         1         2         3         4         5         6         7
0.3489107 0.2389611 0.2380363 0.3753805 0.6617769 0.4456561 0.3185074

            

      
      
      
    

        
What do you meant when you write "it does not work as a normal list"?  
Admittedly it is a name-less list, which seems a bit non-traditional,  
but accessing as avgs[[1]] or avgs[[2]] gives appropriate values. (And  
_you_ are the one responsible for creating this nameless object.)  If  
you just look at the beginning of the str results:

 > str(avgs)
List of 20
  $ :Classes 'summaryDefault', 'table'  Named chr [1:2] "logical" "1"
   .. ..- attr(*, "names")= chr [1:2] "Mode" "NA's"
  $ :List of 7
   ..$ si.summary     :Classes 'summaryDefault', 'table'  Named num  
[1:6] -0.435 0.561 0.657 0.516 0.687 ...
   .. .. ..- attr(*, "names")= chr [1:6] "Min." "1st Qu." "Median"  
"Mean" ...
<snipped remainder of output>

You can then access the first leaf of the second element:

avgs[[2]]$si.summary
    Min. 1st Qu.  Median    Mean 3rd Qu.    Max.
-0.4349  0.5615  0.6574  0.5160  0.6867  0.7525

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      jim holtman
    

    

      
      #
    

  


  

      
Is this what you want:

            

      
      
      
    

        
+ summary(silhouette(cutree(d,x),
+
+ dist(iris[,-5]))))

            

      
      
      
    

        
+     cat('i=', i, 'average=', avgs[[i]]$avg.width, '\n')
+ }
i= 2 average= 0.515983
i= 3 average= 0.5135953
i= 4 average= 0.4998128
i= 5 average= 0.346174
i= 6 average= 0.3382031
i= 7 average= 0.3297649
i= 8 average= 0.324025
i= 9 average= 0.3191681
i= 10 average= 0.3028503
i= 11 average= 0.3072648
i= 12 average= 0.2834498
i= 13 average= 0.2776717
i= 14 average= 0.2855396
i= 15 average= 0.2745142
i= 16 average= 0.2578903
i= 17 average= 0.2531909
i= 18 average= 0.2473504
i= 19 average= 0.2484205
i= 20 average= 0.2545357

            

      
      
      
    

        

        
On Fri, Dec 31, 2010 at 10:34 AM, ADias <diasandre at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      ADias
    

    

      
      #
    

  


  

      
Hello,

thank you all. I have been able to solve my problem with your help.

The problem I am trying to solve is:

I am working on a clustering method to group a data base. At the moment I am
using the clustering hierarchical method and trying to get to the best K
group value via the silhouette function. 

My database has 569 observations and the dendogram is not readable as you
can see on the picture attached. So I am trying to find a way to get to a
conclusion without the use of the dendogram.

thanks again,
Regards,
A.Dias

http://r.789695.n4.nabble.com/file/n3169962/dendogram.jpeg