Actions

Difference between revisions of "Scaling in Biological and Social Networks - Abstract - Clauset"

From Santa Fe Institute Events Wiki

 
Line 5: Line 5:
 
Aaron Clauset (Santa Fe Institute)
 
Aaron Clauset (Santa Fe Institute)
  
Abstract to be posted shortly.
+
This talk will be in two parts. In the first half, I will describe the general task of extracting the "large-scale structure" of networks, giving a brief taxonomy of the existing methods, their advantages and disadvantages, and discussing what exactly we mean by "large-scale structure." In the second part of the talk, I'll describe in more detail a model-based approach to extracting and characterizing the hierarchical structure of networks. I'll also briefly describe the idea of fitting models to data using the principle of [http://en.wikipedia.org/wiki/Maximum_likelihood maximum likelihood], as a warm-up for fitting our generative hierarchical model, called a hierarchical random graph (HRG), to data. Finally, I'll briefly show that hierarchy can explain many of the network statistics more commonly measured, and can predict missing struture in networks.
 +
 
 +
Further reading:<br/>
 +
[http://arxiv.org/abs/physics/0610051 Structural Inference of Hierarchies in Networks], A. Clauset, C. Moore and M.E.J. Newman (2006). <br/>
 +
[http://arxiv.org/abs/0706.1062 Power-law distributions in empirical data], A. Clauset, C.R. Shalizi and M.E.J. Newman (2007).

Latest revision as of 18:10, 10 July 2007

Workshop Navigation

Learning the Large Scale Structure of Networks

Aaron Clauset (Santa Fe Institute)

This talk will be in two parts. In the first half, I will describe the general task of extracting the "large-scale structure" of networks, giving a brief taxonomy of the existing methods, their advantages and disadvantages, and discussing what exactly we mean by "large-scale structure." In the second part of the talk, I'll describe in more detail a model-based approach to extracting and characterizing the hierarchical structure of networks. I'll also briefly describe the idea of fitting models to data using the principle of maximum likelihood, as a warm-up for fitting our generative hierarchical model, called a hierarchical random graph (HRG), to data. Finally, I'll briefly show that hierarchy can explain many of the network statistics more commonly measured, and can predict missing struture in networks.

Further reading:
Structural Inference of Hierarchies in Networks, A. Clauset, C. Moore and M.E.J. Newman (2006).
Power-law distributions in empirical data, A. Clauset, C.R. Shalizi and M.E.J. Newman (2007).