Introduction to Network Modeling

• Activity Number: 258
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Alan Izenman*
• Companies: Temple University
• Keywords: community detection ; random graphs ; stochastic blockmodel ; degree distribution ; configuration model ; ERGM

Abstract

The modeling of networks has become a very interesting and important part of understanding how networks form and how they grow. In this talk, we discuss the modeling of different kinds of networks, such as the Internet and the World Wide Web, financial networks, social networks, telephone networks, transportation networks, and biological networks. Each of these networks has its own peculiar structure, and interest focuses on detecting that structure and creating a suitable model to explain that structure. Simple networks can often be modeled as random graphs. For more highly structured networks, we describe the problem of community detection, which leads to stochastic blockmodels. We discuss some of the defining characteristics of a network, such as the appearance of a giant component, the power law and scale-free networks, and degree distribution. A popular model that specifies a degree distribution of a network is the configuration model. Other models, such as preferential attachment and random node copying, have been proposed to explain the growth of networks. Several parametric network models, such as exponential random graph models, have also been proposed.