| Online social networks like Twitter and Facebook are playing important roles in people's daily life nowadays. Due to their large user base, ubiquitous access through different classes of devices, easy production of content and high traffic, they are becoming ideal platforms for information propagation, both benign and malicious, with the latter leading to serious security issues. To better understand how information propagates on these networks, it is important to study their structure and formation process. In this thesis, the structure and formation process of Twitter are studied to understand the issue relating to its role as a news media or another social network. It has been shown in the literature that the degree distribution of Twitter structure does not obey a power law distribution, as observed in many other online social networks. To address the issue relating to its role two large empirical datasets containing the whole topology of Twitter network have been analyzed, and the analyses suggests that Twitter structure has a component network following power law distribution, with the power law exponent similar to some other examples. Also, based on the analyses, we infer that different components of the Twitter network can be mapped to different roles in information propagation. The thesis then proposes a concise configurable model that can generate a network similar to the Twitter network in two steps. The model formation and its validity is verified by mathematical analysis as well as large scale simulation. The potential of extending this model to generate other online social networks and the impact as well as security implications of the proposed structure on information propagation has also been discussed. |
