2
$\begingroup$

I was looking a paper, article, book... about MCMC (Markov Chain Monte Carlo) methods applied to networks (networks of people, computers,...)

I haven't found something "useful", but maybe I am making this up and I hasn't been applied.

So, do you know some resources where I can look?

$\endgroup$
  • 2
    $\begingroup$ Can you say more about what you want from such a reference? $\endgroup$ – gung Nov 27 '16 at 23:09
  • $\begingroup$ Information, I want to make a brief presentation about it, @gung $\endgroup$ – user140214 Nov 28 '16 at 10:08
  • $\begingroup$ "Information" isn't very specific. If you're doing a presentation, are you just looking for a quickie overview? $\endgroup$ – gung Nov 28 '16 at 12:30
  • $\begingroup$ Maybe a quick overview with some important results, because it will help if I can make a demonstration of the prove of the results and/or an implementation of some algorithm $\endgroup$ – user140214 Nov 28 '16 at 19:06
2
$\begingroup$

MCMC can be used for social network analysis and also for sampling from a network. Here is a short description of the way it is used in the two capacities.

Social Network Analysis: Whenever social networks are analyzed using Bayesian methodology, it is likely that MCMC will be used. One example is Latent Space Models. You might be able to find more references therein.

Sampling from a Network: Often the network is too large to handle, or the network represents a hard to reach population. In this case, a sample from the network is obtained to make inference on the population. Standard sampling techniques often don't work since the population might be unknown. In this case, MCMC crawling techniques may be used. You should be able to find more references therein.

$\endgroup$
  • 1
    $\begingroup$ +1 Excellent references. One additional note, SNA is a relatively young methodology and, as such, has yet to integrate many of the developments from frequentist, Bayesian, time series, etc., methods. A good example of this is that classic time series concerns such as autocorrelation, stationarity, unit roots, cointegration, etc., have not seen application in longitudinal SNA. This is testimony to its relative recency. Same thing with Bayesian methods as applied to SNAs. $\endgroup$ – DJohnson Nov 30 '16 at 14:33
  • $\begingroup$ I comment a bit late here, but do you know more examples of MCMC applied to social network analysis? $\endgroup$ – user140214 Dec 30 '16 at 10:50
  • $\begingroup$ @rdguez Not really. Like I said, references within the papers I mentioned might be able to help you out. $\endgroup$ – Greenparker Dec 30 '16 at 18:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy