Overview / Usage

Check out our video: https://www.youtube.com/watch?v=Mmonc7ox8sc&feature=youtu.be

Social networks disseminate real-time information through users’ dynamic interactions and collective behaviors. Consequently, social networks manifest remarkable bursts of keywords and topics that correspond to real-world events that draw intensive attention from the general public. One of the main motivations behind social network analysis is the quest for understanding opinion formation and diffusion.

To harvest the full potential of this new medium of information, our group focuses on developing methods that discover complex interactions and collective behaviors that determine how various types of bursty events in social networks (e.g., intensive attention of the general public) are generated and propagated.

The results of this research can be used to promote or to impede the propagation of crucial information within social networks. Further, our group focuses on designing early detection and forecasting engines for bursty events that take place over large social networks. The results of this research represent an important step toward real-time and predictive management of social networks by making possible a response in space and time that is valuable for various applications, such as micro-blogging, viral marketing, emergency response, and disaster management.

Methodology / Approach

In order to build a model that properly accounts for the existence of distinct opinion formation phases and social balancing, we intend to employ discrete event simulation on diverse social network topologies to show the importance of the network size and opinion source ratio. We also examine the phase transition to social balancing fostered by the democratic structure of the small-world topology. Last, we incorporate a dynamics engineering framework; a deep learning model; and a data-driven evaluation method in place of the expensive field experiments.

Other links

• SLD: System Level Design Group