On Mon, Jan 07, 2008 at 05:25:44PM +0100, Lorenzo Isella wrote:
> Thanks for both the replies.
> I am now giving a try to the suggestion by Gabor since it looks easier
> (for me) to implement.
> I am testing it, but so far it does what I have in mind.
> I am going now through the documentation of t
Thanks for both the replies.
I am now giving a try to the suggestion by Gabor since it looks easier
(for me) to implement.
I am testing it, but so far it does what I have in mind.
I am going now through the documentation of the igraph package. I can
count the cluster number, but I also want to make
Dear Lorenzo,
if I understand your posting correctly, this is exactly what Single
Linkage clustering does if you cut the dendrogram tree at your threshold
distance.
Therefore you can use hclust with method="single" (which produces the full
dendrogram; you have to generate the Euclidean distanc
Lorenzo, why can't you actually generate the graph to find the
connection components? With the 'igraph' package this is something like:
g <- graph.adjacency( DIST < 0.5, mode="undirected" )
g <- simplify(g)
no.clusters(g)
assuming you have your distance matrix in 'DIST'. If N is too big
then you
Dear All,
I hope I am not asking a FAQ. I am dealing with a problem of graph
theory [connected components in a non-directed graph] and I do not
want to rediscover the wheel.
I saw a large number of R packages dealing for instance with the
k-means method or hierarchical clustering for spatially dist
5 matches
Mail list logo
