[R] Dissimilarity matrix and number clusters determination

[Luisfo Chiroque]

Dear Michael,

Yes, AFAIK you are correctly reading the results.
You can print
elbow.obj$k
to obtain the optimal number of clusters, and ‘visually’ you can check it plotting the variance vs #clusters
plot(css.obj$k, css.obj$ev)

HTH

Best,
Luisfo Chiroque
PhD Student
IMDEA Networks Institute
http://fourier.networks.imdea.org/people/~luis_nunez/ <http://fourier.networks.imdea.org/people/~luis_nunez/>
> El 12 abr 2016, a las 4:30, Michael Artz <michaeleartz at gmail.com> escribió:
> 
> Hi,
>  I already have a dissimilarity matrix and I am submitting the results to
> the elbow.obj method to get an optimal number of clusters.  Am I reading
> the below output correctly that I should have 17 clusters?
> 
> code:
> top150 <- sampleset[1:150,]
> {cluster1 <- daisy(top150
>                   , metric = c("gower")
>                   , stand = TRUE
>                   , type = list(symm = 1))
> }
> 
> dist.obj <- dist(cluster1)
> hclust.obj <- hclust(dist.obj)
> css.obj <- css.hclust(dist.obj,hclust.obj)
> elbow.obj <- elbow.batch(css.obj)
> 
> [1] "A \"good\" k=17 (EV=0.80) is detected when the EV is no less than
> 0.8\nand the increment of EV is no more than 0.01 for a bigger k.\n"
> attr(,"class")
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.


	[[alternative HTML version deleted]]