Estimating epidemic spread including social media

Official URL: http://risc01.sabanciuniv.edu/record=b1655721 (Table of Contents)

Epidemic growth models are usually designed by considering only the infection status of people. However, information about the epidemic has a significant effect on individuals’ behaviors with the increasing use of social media. When the epidemic starts spreading, the information about epidemic also spreads among people. In this thesis, we provide a new deterministic model, namely UVIW (Uninformed-susceptible, Informed-susceptible, Infected, Social Media) which incorporates the epidemic spread as well as information spread. This model is designed by considering the SIS (Susceptible - Infected – Susceptible) model as a base model [1]. We divide the susceptible state into two new states based on the awareness status of people about the diseases and include the social media state to calibrate the dynamics of social media. Furthermore, we use Influenza like illness data of USA and related twitter data to analyze the predictions of our model, and we compared the results with logistic growth model. We indicate that using the model with information spread yields better estimations especially in the early growth phase of the epidemic. Additionally, our model produces predictions for the social media dynamics which is also a novel outcome in this field.