* Faculty       * Staff       * Contact       * Institute Directory
* Undergraduate       * Graduate       * Institute Admissions: Undergraduate | Graduate      
* Events       * Institute Events      
* Lab Manual       * Institute Computing      
No Menu Selected

* Research

Ph.D. Theses

Agent-based Dynamics Models for Opinion Spreading and Community Detection in Large-scale Social Networks

By Jierui Xie
Advisor: Boleslaw Szymanski and Gyorgy Korniss
May 11, 2012

In this thesis, two topics in the context of social network analysis (SNA) are studied. One is the opinion dynamics, and the other is the community structure discovery.

In order to provide useful insights into understanding the evolution of opinions, ideologies or attitudes, this thesis explores a simple, abstract opinion dynamics model called binary agreement model, where there are two competing opinions. The contribution of this thesis is to quantify the effect of committed minorities who hold unshakable opinions. In particular, such effect is investigated in two scenarios: (1) The first scenario is one committed group competing with the opposing uncommitted majority (2) The second scenario is the more general case where two groups committed to distinct opinions A and B, and constituting fractions pA and pB, coexist. As a comparison, the influence of global social media is briefly discussed, which serves as a kind of committed opinion.

Mining communities that allow multiple memberships is challenging especially in large-scale networks. This thesis presents a fast algorithm called SLPA for overlapping community detection, which is successfully applied to networks with million nodes. Detecting and tracking communities in a dynamic network where changes arrive as a stream is another challenging issue in real-world applications (e.g., real time monitoring of Internet traffic or online social interactions). Instead of computing communities on each snapshot independently, this thesis proposes LabelRankT, a decentralized online algorithm, to detect communities in large-scale dynamic networks.

* Return to main PhD Theses page