Formation and Evolution of Groups in Large Social Networks
Daniel P. Huttenlocher
Mauldin Auditorium (NSH 1305 )
Talk 3:30 pm
This talk will consider factors that influence people
to join communities, and in turn how communities develop and evolve over time.
While there is a long history of research on such questions in the social
sciences, recent online records of social interactions offer an unprecedented
ability to study these issues at a large scale. We have been
investigating questions such as what properties of social networks influence
whether individuals will join communities, how the influence of social ties
compares with other factors such as shared interests, and what causes
communities to grow more or less rapidly. We are studying these questions using
several sources of data including LiveJournal, Wikipedia and DBLP, each of
which provides both social ties and explicit user-defined communities. We
find surprising consistency across datasets in terms of how a person's
propensity to join a community depends on the number of social ties to that
community. In contrast, the relative importance of social ties versus
shared interests seems to depend considerably on the setting.
This is joint work with Lars Backstrom, Dan Cosley, David Crandall, Jon Kleinberg, Xiangyang Lan and Sid Suri.
Dan Huttenlocher is the John P. and Rilla Neafsey
Professor of Computing, Information Science and Business at
For appointments, please contact Alexei Efros (email@example.com)