Speaker: Mark Handcock
Title: Assessing the Goodness-of-Fit of Network Models
Network models are widely used to represent relational
information among interacting units and the implications of
these relations. In studies of social networks recent emphasis
has been placed on random graph models where the nodes usually
represent individual social actors and the edges represent a
specified relationship between the actors. Much progress has been made on
exponential family models for cross-sectional networks, and some has been
made
on related models for networks observed longitudinally.
We present methods to assess how well estimates of network structure fit the
network data on which they are based.
The first such methods are based on the likelihood-ratio statistics.
The second compare structural statistics of the observed data
with the corresponding statistics on graphs
generated from the fitted model.
We apply this approach to the
study of friendship relations among high school students from the
National Longitudinal Study of Adolescent Health. The
sizes of the networks we fit range from 71 to 2209 nodes.
This is joint work with David R. Hunter and Steven M. Goodreau.