Diffusion in Social Networks by Soham Das
Posted on January 14, 2016
The main aim of organizing the Grad Talk is to foster research interest in the young students in our Chapter who have finished their high school and are currently pursuing undergraduate studies. This also gives the opportunity to the members of the Student Chapter for a face-to-face interaction with grad students a few years ahead of us along the academic path and exposure to the work being done by them.
26th July 2012, 3:15 p.m.
Speaker: Soham Das
Title: A Study on Topic Diffusion in Social Networks (relevant paper link on arxiv)
Abstract: Social Networks have become an integral part of our life today, but still, very little is known about how some mundane and ordinary topics become viral, while others die out fast in these networks. Here, we propose a competitive stochastic framework for modelling the diffusion of topics in a social network. The underlying topology in our model is a small world network, based on the Watts & Strogatz Graph. When instantiated properly, our model attains the viral regime, closely mimicking the behaviour of real-world social networks. Extensive simulation studies have been carried out to explore the parameter space of the model and observe its behaviour under different instantiations, which revealed some interesting trends. The most significant finding in this study is the phenomenon of lattice cluster merging, by which clusters of proximate users talking on the same topic (we calllattice clusters), expand on both sides and finally merge along the lattice to form bigger lattice clusters. These big lattice clusters can sustain the topic for a longer period of time and eventually help it to go viral. This implies that it is not the weak links but actually the strong links of the Watts & Strogatz Graph, that play a major role in the virality of the topic. The fact, that network dynamics dominate in driving a topic viral, is further strengthened by our observations, which show that the adopters hardly play any role in the topic’s virality.
About the Speaker:
I did my B.Tech in Computer Science from IEM, Kolkata in 2010. I have just completed my M.Tech in Computer Science from IIT Delhi. I worked on Modeling and Analyzing Social Networks for my M.Tech dissertation. I will be joining University of Florida, Gainesville for my PhD this coming August. My primary research interests are Network Algorithms, Social Networks & Data Mining.