Hello
I have a large dataset to make a cluster analysis and I'm trying to use
Rclusterpp to do so.
Nevertheless in a sample matrix analysis (see
http://dl.dropbox.com/u/755659/gillnet.txt) I'm getting different results
from Rclusterpp and hclust.
Following the example from "An Introduction to Rclusterpp" (
http://cran.r-project.org/web/packages/Rclusterpp/vignettes/Rclusterpp.pdf)
I got:
dat.gillnet <- read.csv("http://dl.dropbox.com/u/755659/gillnet.txt
",sep=",",dec=".",row.names=1)
h <- hclust(dist(dat.gillnet, method="euclidean"),
method="ward")
r <- Rclusterpp.hclust(dat.gillnet, method="ward",
distance="euclidean")
identical(h$merge, r$merge) && all.equal(h$height, r$height)
[1] FALSE
If one plot the clusters will see the differences.
But the purpose of the package is to "deliver identical results as the
“stock” stats::hclust implementation."
Am I doing something wrong? Thanks for any help.
Best regards,
Antonio Olinto
[[alternative HTML version deleted]]
