[gov3009-l] Blitzstein on "Stability and Estimability of Centrality in Networks"
mblackwell at iq.harvard.edu
Tue Mar 23 16:41:44 EDT 2010
We hope you will join us this Wednesday, March 24th at the Applied
Statistics workshop when we will be happy to have Joe Blitzstein
(Department of Statistics). Details and an abstract are below. A light
lunch will be served. Thanks!
"Stability and Estimability of Centrality in Networks"
Harvard Department of Statistics
March 24th, 2010, 12 noon
K354 CGIS Knafel (1737 Cambridge St)
"Centrality" is one of the most widely-used notions network analysis,
attempting to measure the importance of a node. Yet despite being such
a central idea to the study of networks, there is no standard
definition; rather, there are many incompatible definitions, e.g.,
based on degrees, shortest paths, and eigenvectors of the adjacency
matrix. We will attempt to sort through this menagerie of definitions
of centrality from the perspective of stability and estimability. That
is, if the network is perturbed slightly or is measured with noise,
how much does centrality change? If -- as is usually the case -- only
a sample of the full network is observed, which of these notions can
be reliably estimated from the sample?
Institute for Quantitative Social Science
Department of Government
email: mblackwell at iq.harvard.edu
More information about the gov3009-l