Title: Analysing Social Communities

Abstract

We analyze 'Social Communities' for learning and predicting their online
behaviors.  Community reactions such as commenting, liking, sharing,
retweeting, responding etc. are observable and explicit. These reactions are
often influenced by the content of the post. However, the reactions are also
affected by the content independent factors such as events occurring external
to the community, geo location of the community, length of the post, author of
the post, etc. There are other community behaviors that are not explicitly
observable. The behaviors such as  formation of sides, subgroups, leaders,
influencers and detractors etc. are not explicitly observable. These implicit
behaviors essentially perturb the social graph in response to a post or
external stimuli. I will briefly describe our research efforts in learning and
predicting these online community behaviors. Our work relies on publicly
available unstructured social data (e.g. facebook, twitter and blogsite data)
and also on general content hosting sites such as Wikipedia and DB Pedia.

I will also provide a brief overview of the Analytics related research work
being done at Adobe India Advanced Technology Labs.

Bio

Dr. Shriram Revankar is the Director of Adobe India Advanced Technology Labs.
Previously he was with Xerox Corporation, where he was a Xerox Fellow. He
served Xerox in various capacities including as the head of Smart Document
Systems Lab (NY, CA); as the Chief architect of Xerox's Production Solutions
and Services business; and as a Principal Scientist at Xerox Webster Research
Center. In his current capacity Shriram is focused on building a new world
class research and technology lab for Adobe Systems in Bangalore, India.
Shriram has a M.S. and Ph.D. in Computer Science from SUNY Buffalo, and his
research interests span smart and adaptive systems, Social Analytics, cloud and
Bigdata; and collaboration services on non-traditional computing platforms