skip to content
 

Edge exchangeability: a new foundation for modeling network data

Presented by: 
Harry Crane Rutgers, The State University of New Jersey
Date: 
Monday 25th July 2016 - 13:30 to 14:00
Venue: 
INI Seminar Room 1
Abstract: 
Exchangeable models for vertex labeled graphs cannot replicate the large sample behaviors of sparsity and power law degree distributions observed in many network datasets. To address this issue, we introduce the principle of edge exchangeability, which is more natural for most applications and admits models for networks with sparse and/or power law structure. The vertices in an edge exchangeable network arrive in size-biased order according to their degree, further explaining why vertex exchangeability is an untenable assumption for many applications.

Our discussion settles a longstanding question of statistical network modeling and presents a new framework within which to develop future theory and methods.

Joint work with Walter Dempsey.
The video for this talk should appear here if JavaScript is enabled.
If it doesn't, something may have gone wrong with our embedded player.
We'll get it fixed as soon as possible.
University of Cambridge Research Councils UK
    Clay Mathematics Institute London Mathematical Society NM Rothschild and Sons