University of Konstanz
Graduiertenkolleg / PhD Program
Computer and Information Science

Graduation Talks

title

Tools for Actor-Based Modeling of Dynamic Networks

speaker

Natalie Indlekofer, University Konstanz
Konstanz, Germany

date & place

Wednesday, 23.06.2010, 16:15 h
Room C 252

abstract

Actor-based modeling is the most prominent approach for statistical analysis of longitudinal social networks. Given a sequence of observations of a network at discrete moments in time, actor-based models are designed for detecting network-specific formation rules, called network effects, that govern the unobserved evolution between consecutive states. This evolution is simulated by a parameterized stochastic process. Each parameter is associated with one effect and estimated to fit the observed data. The estimates indicate the influence of corresponding effects on network dynamics.
However, estimated parameters are hardly interpretable, since they are not comparable across different data or effects. Evaluating whether a model represents the observed data sufficiently is indispensable for reliable analysis. The aim is to provide systematic guidance on model improvement and interpretation.