Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: clustering sequences - problem with cluster and clustermat


From   Nick Cox <njcoxstata@gmail.com>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   Re: st: clustering sequences - problem with cluster and clustermat
Date   Tue, 16 Apr 2013 11:05:25 +0100

I don't know what your first sentence means.

But my guess is that there isn't enough memory for one of the programs
called by -cluster- to create a matrix.

Precisely what the memory limit is clearly of interest to you, but I
can't see that it's your main problem. Your main problem is, I
suggest, making your goal here clearer to yourself.

2700 observations is a _big_ classification problem, particularly when
you consider that the number of comparisons goes up with the square of
the number of objects.

Even more crucially, clustering a single variable, which you are
trying to do, can't show much -- arguably, can't show anything --
that's not evident by looking for clusters on a distribution display.
With 2700 observations, a histogram might be too coarse and you will
get better detail from a quantile plot or a dot plot; but what's
crucial is how many distinct values there are, not the number of
observations.

If there are gaps, and thus clusters, they will be obvious on a
sufficiently detailed graph. Otherwise your variable is a continuum.

I can't improve on what is said here by Stata user Phil Ender

http://www.philender.com/courses/multivariate/notes2/cluster0.html

Nick
njcoxstata@gmail.com

On 16 April 2013 10:36, BUSSI, Margherita <mbussi@etui.org> wrote:

> I've done a sequence analysis based with subcost based on the means of transitions between states.
> I've got a new variable that it not a dissimilarity matrix (as I wrongly? expected)
> I tried to operate a cluster analysis but I got this message:
>
> Cluster medianlinkage meanprob
> cluster name: _clus_1
> insufficient memory for  ClusterMatrix
>
> What memory is stata talking about? Storing? Running memory? And why Clustermatrix if when I give the command "clustermat" it does not find the matrix?
>
> if I ask
> . cluster  singlelinkage meanprob
> It runs forever...
> (I've 3700 observations)
>
> Does anyone have an idea how to solve it or where is my mistake?
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index