Let me quote the mail I received:
I'm glad I've found your excellent clustering package
However, one thing I find not intuitive:
>>> from cluster import *
>>> data = [12,34,23,32,46,96,13]
>>> cl = HierarchicalClustering(data, lambda x,y: abs(x-y))
>>> cl.getlevel(10) # get clusters of items closer than 10
[96, 46, [12, 13, 23, 34, 32]]
Here I would expect every item in the list being a list itself, like
[, , [12, 13, 23, 34, 32]]
Normally this should not cause many problems, as python (at least so it seems to me) handles this correctly when you do a "for cluster in cl.getlevel(10)" for example.
However, I admit that this is something that should be handled more consistently internally by the module. This will improve overall code readability when using the module.
Log in to post a comment.