Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
st: comparing matrices generated through cluster analysis
From
Rachel Lindenberg <[email protected]>
To
[email protected]
Subject
st: comparing matrices generated through cluster analysis
Date
Tue, 6 Jul 2010 15:34:29 -0700
I am currently conducting research using cluster analysis. First I
classified/clustered survey respondents based on whether they belong
to certain types of groups i.e. the variables below called Civil Moral
Cultural Political Economic are binary:
cluster averagelinkage Civil Moral Cultural Political Economic,
measure(matching)
I know how to get diferent cut points
cluster gen alife9 = groups(9)
cluster gen alife3 = groups(3)
I think I understand how to analyze which result of clusters provides
the best "fit" to the data. So far, I have used things like:
cluster stop alife, rule(cal)
But what I want to do now is assess the fit of clustering based on the
use of different resemblance coefficients, different clustering
methods or even different input variables. I want to be sure I am
using the best method and my final goal is to assess whether clusters
generated by group activity (as above) correspond to those generated
by income and eduation. I understand conceptually that what is
required is to unstring the resemblance matrix (or in some cases the
cophenetic matrix) and find a matrix correlation coefficient. My
problem is I don't understand how to do this in STATA. I'm having
trouble understanding how STATA commands correspond to these
conceptual ideas and how to save/work with the matrices
Does anyone have or know of a sample program that clearly explains the
appropriate procedure for this kind of analysis of even this final
part?
Thanks,
Rachel
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/