MS Teams (please email firstname.lastname@example.org for the meeting link)
Zitao He | Applied Mathematics, University of Waterloo
Applications of Machine Learning and Other Data-driven Methods in Epidemiology
Mechanistic models and machine learning are two different types of modelling methods—the former method is developed through a deductive process and focuses on the underlying physical laws, while the latter uses inductive reasoning and attempts to build a direct connection between data. Problems in epidemiology have long been studied with mechanistic models as they successfully explain why and how individuals progress between compartments, changing the populations in size. These days, with a drop in the cost of computation and increasing ease of collecting data, machine learning's strength in tackling massive multiscale data has been amplified. Recent advances in machine learning algorithms have made it a reliable predictive tool, even though it could provide non-physical solutions. This thesis aims to explore synergic ways by incorporating the advantages of mechanistic models and machine learning, addressing problems in the field of epidemiology related to disease prediction, prevention, and control.