We had this question a couple of weeks ago (see the archive). No, there is currently no function for bootstrap support indices on hclust objects. However, code contributions are welcome. ;-) A long-winded way to do it could be to convert your hclust objects to phylo objects, using as.phylo (package ape), then write them to a file as bracket-format trees using write.tree (package ape), then analyse the results in another package e.g. PHYLIP.
Simon. Simon Blomberg, PhD Depression & Anxiety Consumer Research Unit Centre for Mental Health Research Australian National University http://www.anu.edu.au/cmhr/ [EMAIL PROTECTED] +61 (2) 6125 3379 > -----Original Message----- > From: Ron Ophir [mailto:[EMAIL PROTECTED] > Sent: Monday, 14 July 2003 5:10 AM > To: [EMAIL PROTECTED] > Subject: [R] bootstrap for hclust > > > dear group members, > I am looking for a function that assess the stability of > cluster. The result of hclust function is an hclust object > which can be plot as a dendrogram. However to have > confidence in the tree topology usualy bootstap is applied. I > understand that I can apply bootstarp on the original data > and then run hclust(dist() ) as much as I resampled but how > to comapre the topologies the I don't know. Is there a > bootstarp or permutation for cluster? > thanks, > Ron > > ______________________________________________ > [EMAIL PROTECTED] mailing list > https://www.stat.math.ethz.ch/mailman/listinfo/r-help > ______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
