***** To join INSNA, visit http://www.insna.org *****
I just heard a talk arguing that spectral comparisons are the best way to
compare networks. This was in the context of privacy-preserving strategies,
so it might not translate well into simulated networks. If there's a direct
mapping from nodes in your empirical data to nodes in your simulated data,
spectral measures might be appropriate.
Xintao Wu, Xiaowei Ying. Randomizing Social Networks: a Spectrum Preserving
Approach. 2008 SIAM Conference on Data Mining.
On Sun, Apr 27, 2008 at 5:54 PM, simone gabbriellini <
[log in to unmask]> wrote:
> ***** To join INSNA, visit http://www.insna.org *****
> Dear List,
> I would like to have your opinion on this methodological issue:
> - suppose you have a network retrived from empirical data;
> - suppose you have simulated 100 networks (with same size and edge
> number), trying to approximate that empirical network (assume that to
> approximate means to minimize the hamming distance between the empirical and
> the simulated networks - or the mean matrix of the simulated stack);
> what will your testing strategy be? what procedures will you suggest me to
> test the significance of the distances between the simulated and the
> empirical networks?
> Best regards,
> Simone Gabbriellini
> dott. Simone Gabbriellini - PhD Student
> Department of Social Sciences
> University of Pisa
> [log in to unmask]
> SOCNET is a service of INSNA, the professional association for social
> network researchers (http://www.insna.org). To unsubscribe, send
> an email message to [log in to unmask] containing the line
> UNSUBSCRIBE SOCNET in the body of the message.
School of Computer Science
Carnegie Mellon University
5000 Forbes Ave.
Pittsburgh, PA 15213
[log in to unmask]
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.insna.org). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.