seeking math ref for multi dimensional scaling

[Judy Brown <brown@xxxxxxxxxxxxx>]

I've been using matlab's function cmdscale for MDS, and am intrigued
by the algorithm ie would like to understand the derivation.  They
reference Seber, G.A.F., Multivariate Observations, Wiley, 1984, which
I've ordered but will take the lib months, and may not have what I
want anyhow.

I know there are thousands of MDS papers out there, but doubt if many
discuss the derivation, and of those probably not matlab's algorithm
(I read Kruskal who talks about linear regression rather than
eigenvalues.)   So if anybody knows a good reference (short and simple
minded preferred), I'd be very grateful.

thanks,
Judy



--
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
    Judy Brown            |http://www.media.mit.edu/~brown
    jbrown @ wellesley.edu|http://www.wellesley.edu/Physics/brown/jbrown.html
    brown @ media.mit.edu |E15 483,   MIT, Cambridge, MA 02139
                          |584 Science Cnt, Wellesley College, 02481
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*