Hi Bill,
actually I said "already classified" but that does not make much
difference. I am getting some intriguing ideas off-list which I will
summarise with thanks later, but I don't want to choke off the flow by
getting too specific or evaluative right now. I'm keen to respond to
your contribution but I won't do that now for the same reason.
Cheers, Murray
On 19/09/2011 9:52 a.m., [log in to unmask] wrote:
> I take it "already clustered" means you have an existing partition of the entities into discrete groups, as opposed to some sort of hierarchical clustering, and you would like the next step to preserve these groups.
>
> Isn't this just a problem of clustering the clusters? In my naivete I would have thought all you need to do is
>
> a) define a distance measure between groups, (as opposed to the primary entities), that reflects the importance you put on the separations that exist already,
>
> b) define an overall objective you want the clustering to achieve (e.g. as an objective function to optimise, that balances costs and benefits) and
>
> c) employ some appropriate clustering algorithm to see if there is any advantage in merging any of your existing groups. Which one to use will depend on your circumstances. (At this stage it might sometimes be useful to look at a hierarchical clustering of your groups first, rather than to go for a hard clustering straightaway - or it may not!)
>
> What have I missed?
>
> Bill.
>
> -----Original Message-----
> From: [log in to unmask] [mailto:[log in to unmask]] On Behalf Of Murray Jorgensen
> Sent: Sunday, 18 September 2011 3:42 PM
> To: [log in to unmask]; anzstat
> Subject: Clustering classified data
>
> I would like to ask the list a question which seems to subtle to yield
> to Google. Suppose we wish to cluster multivariate data that are already
> classified. Certainly data in the same group of the classification
> should be in the same cluster but we may like to find a coarser
> clustering in which some of the classification groups are merged.
>
> I can think of a few things that might be done but I wonder if anyone
> can point me to papers on such questions - theoretical or applied?
>
> Cheers, Murray
--
Dr Murray Jorgensen http://www.stats.waikato.ac.nz/Staff/maj.html
Department of Statistics, University of Waikato, Hamilton, New Zealand
Email: [log in to unmask] Fax 7 838 4155
Phone +64 7 838 4773 wk Home +64 7 825 0441 Mobile 021 0200 8350
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|