Dynamic Tie Strength Structure in Online Social Networks

With the rapid diffusion of online social networking services, human social interactions have not only been augmented but also become digitally documented. With more and more people shifting their primary social networking activities online, companies are eagerly exploiting the unprecedentedly rich social networking data and harnessing the power of online social networks. However, theoretical understanding about online social connections is limited and most existing studies treat online connections as homogenous binary conduits of influence. Recognizing that online social connections are multi-relational, heterogeneous and dynamic, this research sets to uncover the hidden dynamic structure of heterogeneous social influence tie strength embedded in online social networks. In this research, based on a comprehensive review of the theoretical literature on social networks, we will introduce both a static model and a dynamic model of tie strength structure in online social networks that allow individual, category and dyad-level heterogeneity; and we will propose efficient estimation procedures to uncover the hidden social influence tie strength from observed user activities and online social connections. The proposed theoretical models and estimation methods could generate estimates that are used to extend social network theories in the context of online social networks. Further, an analytical application will be built, which is capable of generating real time social network analytics in online social networking service websites.