Quantifying User Attitude in Social Networks
With the proliferation of social networks and their extent of reach (both positive and negative), a large body of research focuses on utilizing and explaining the significance of their impact. One of the critical problems investigated is understanding the information diffusion/influence propagation in social networks. Diffusion refers to the (probabilistic) behavior of the interaction between the entities in the network describing when/how its neighbors' actions influence an entity. Domingos and Richardson's seminal works, and Kempe et al. proposed two popular models for information diffusion-Independent Cascade and Linear Threshold. In these models, a network node is influenced if it receives the information originated at the seed set. We study these basic models and extend them by introducing the strength of influence, i.e., attitude, and develop approximation algorithms for related maximization problems.
Committee: Pavan Aduri (major professor), Samik Basu (major professor), Jia Liu, Qi Li, and Shawn Dorius