Bayesian Mixed-Membership Models of Complex and Evolving Networks
Abstract
This thesis provides a methodological framework for the statistical analysis of complex graphs and dynamic networks.1 In it, I develop probabilistic algorithms that generate, evolve and integrate a heterogeneous collection of graphs, I study the statistical models these algorithms implicitly specify, and I develop strategies for estimating the set of quantities on which they depend in the context of applications to social and biological networks.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2006
- Accession Number
- ADA488405
Entities
People
- Edoardo Airoldi
Organizations
- Carnegie Mellon University