Information diffusion on Twitter
Date of Award
M.S. in Computer Science
Department of Computer Science
Advisor: Zhongmei Yao
With the wide and increasing application areas of social network, both business and academia pay great attention on it. Studies on information diffusion are one of the emerging but still understudied areas. Most of the studies focus on how to leverage the link structure of social network to maximize information diffusion and base their algorithms development and verification on virtual social network structures. As such, researchers may have limited information diffusion within a certain boundary. Furthermore, the application of the algorithms and findings could be weak when facing the real social network structure. To address the gaps of literature, this thesis investigates the impact of single information source and multiple information sources on information diffusion patterns with real world public social network structure, which is collected by designing a functional and ready-to-use data crawling structure. To answer the research question, a simulation experiment was designed and executed. With the simulation data, ANOVA and ordinary least square regress were applied. According to the analysis results, I found that information diffusion patterns vary significantly between single information source and multiple information sources. Specifically, the number of individuals holding the information and the number of individuals retweeting the information, information diffusion pattern parameters based on a variant of susceptible-infected-recovered (SIR) model, increase when there are multiple information sources. The result has significant implications for both academic and managerial area. Specifically, from academic perspective, by identifying the differences of information diffusion patterns resulting from attributes of information source, it would facilitate the researchers to efficiently choose information source detection techniques. From managerial perspective, applicants could design their information source/sources based on their desired information diffusion pattern.
Twitter, Online social networks, Communication Network analysis, Computer Science, Social Network, Web Crawling, Information Diffusion, Simulation, ANOVA, OLS
Copyright 2015, author
Zhou, Li, "Information diffusion on Twitter" (2015). Graduate Theses and Dissertations. 1051.