UofA Team won the Forecasting 2022 Best Paper Award

Graphic Abstract

To help model the course of an epidemic, epidemiologists often use the SIR model, which forecasts the number of people who are Susceptible to the disease, vs Infected, vs Removed, at each future time. This approach requires estimating various parameters – parameters that can change over time, based on changes to public policy, etc.

We describe a way to use machine learning (ML), within the SIR model, to produce the siMLr system. Our empirical analysis demonstrated that siMLr learned models produced state-of-the-art forecasts, and could be interpreted.

The resulting paper "SIMLR: Machine Learning inside the SIR Model for COVID-19 Forecasting" just won the Forecasting 2022 Best Paper Award.

Certificate