Tutorial Talk: Behavior-driven Social Network Mining and Analysis
DASFAA, April 22-25, 2013, Wuhan, China
Ee-Peng LIM, Feida ZHU, Freddy CHUA [
Slides]
Abstract: A distinct feature characterizing social network data is the rich user behavior
information presented most prominently as the various interpersonal social interactions. These
interactive behavior information, often encoded into attributed directed edges, could collectively
reveal tremendous insight into a wide range of social network patterns. In fact, incorporating
these behavior information into the modeling and mining process is sometimes critical for
discovering knowledge at group and community levels. In this tutorial, we discuss how
longitudinal behavior data collected over a time span can be leveraged to facilitate the mining of
a number of interesting social network patterns including virality modeling, anomaly detection,
off-line community identification and behavior-driven topic modeling.