***** To join INSNA, visit http://www.sfu.ca/~insna/ *****
Actually, Ucinet and Netdraw already have transformations of this type,
although crude in netdraw and undocumented in ucinet. In UCINET, if you go
to matrix algebra, you can type:
adjustedA = congru(B A)
This will create a new matrix called adjustedA from a pair of input matrices
A and B. If A and B are coordinate matrices (i.e., rows are nodes, columns
are dimensions, and Xij gives coordinate of node i for jth dimension), then
the program rotates and stretches configuration A to maximally resemble B,
and saves the adjusted A matrix as the output of the routine.
In NEtdraw, the Ties window has a button called More. If you press that you
get a new window containing each relation. You can step through each
relation one at a time, like frames in a movie. The nodes can remain fixed
in position, or move, depending on whether which option you click. If they
move, then a procrustean type computation is performed after the spring
embedding in order to orient each configuration as much as possible to the
previous one. It's crude but may be useful with over-time data.
----- Original Message -----
From: "Balazs Vedres" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Tuesday, February 11, 2003 5:07 PM
Subject: Procrustean analysis and visualizing network dyanmics?
> ***** To join INSNA, visit http://www.sfu.ca/~insna/ *****
> Dear SOCNETters,
> I wonder how SNA packages will move to finally embrace and serve growing
> interest in network dynamics. Especially in visualization I get little
> help from the otherwise (more and more) terrific packages:
> UCINET, NETDRAW and PAJEK.
> A first step that I would find very useful would be to incorporate
> procrustean analysis (or adjustment) into multidimensional scaling (or
> spring embedder) modules that are part of each of these packages.
> Procrustean adjustment would enable those interested in drawing networks
> (of more or less the same actors) in consecutive time periods to see the
> effect of network change on graphs, and filter out the effect of
> contingency (by invariant transformations).
> Is there any package that is able to accomplish this?
> I know that there are some now pretty antique MDS packages that do
> procrustean analysis. Would UCINET, NETDRAW, PAJEK or other packages
> move this way? Or is there some other way?
> Balazs Vedres
> Balazs Vedres
> Ph D student, Sociology
> Columbia University
> mailto:[log in to unmask]
> office ph.: (+212) 854-6591
> office fax: (+212) 854-2963
> home ph.: (+212) 666-3170
> SOCNET is a service of INSNA, the professional association for social
> network researchers (http://www.sfu.ca/~insna/). To unsubscribe, send
> an email message to [log in to unmask] containing the line
> UNSUBSCRIBE SOCNET in the body of the message.
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.sfu.ca/~insna/). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.