Modelling the spread of information and rumours in online social networks

The most retweeted tweet in the world has 5.19 Million retweets. The most liked photo on instagram has 52.2 million likes. The reach and expanse of social networks makes it possible for a piece of information, be it authentic or fake, to travel across the globe and reach millions of people in a matter of seconds. This leads us to the question- what makes a post go viral in an online social network platform? Why are some posts ignored while some make headlines? Starting with these questions, we are modelling information diffusion and rumour propagation in online social networks.

Through our models we aim to compare the results generated by the standard Susceptible - Infected - Recovered (SIR) CA model with the inclusion of population dynamics. Unlike the grid like structure of a conventional SIR model, we are considering a connected, undirected and finite network with N nodes where the nodes represent users in the network and the edges denote the social connections they have. Since real social networks are constantly evolving with people leaving and joining them, we will be considering the effects of new registered users’ growth rate and inactive user rate over time while further building a continuous time model. In addition, we will be comparing the results across three different networks to showcase how changes in the network structure can affect the propagation of information.