Online social networks and other networks of interest are known to exhibit community structure, where a community is defined to be a highly interconnected group of nodes with possibly shared traits or features. However, classic network models, such as the preferential attachment model, do not account for community structure. In this talk, I will present the Community-Aware Preferential Attachment Model (CAPAM), which allows the user to specify community structure via edge probabilities. I will show that CAPAM retains desirable properties of the preferential attachment model, namely a power-law degree distribution, and further that the multivariate degree distribution is dependent upon the edge probabilities in an interesting way. I will show that community structure also plays a role in epidemic spreading processes. Under the SIS model, the lifetime of a spreading process is constrained by the structure of the individual communities, and the epidemic threshold is bounded closely around the threshold associated with the strongest community.
- This event has passed.