Abstract: We present covid19-projections.com, a simple COVID-19 prediction model that adds the power of machine learning on top of a classic infectious disease SEIR model to make forecasts for all 50 US states and more than 70 countries. In this talk, we will explain how we combined the two disciplines to create an effective model that can quickly adapt to new data. Unlike other models that use a multitude of different data sources, we only use past reported deaths to predict future reported deaths. We will also provide insights into what methods worked well and what methods did not, as well as tips for how a data scientist can begin building their own forecasting models. Finally, we give an outlook into where the US may be headed as we head towards fall.
Bio: Youyang Gu is an independent data scientist and the creator of covid19-projections.com. His site has been widely cited by various organizations, including the CDC, to help inform public health decision making. Prior to his COVID-19 modeling work, Youyang honed his research skills as a quantitative researcher in trading and finance. His expertise is in using machine learning to understand data, separate signal from the noise, and make accurate predictions. Youyang completed his Bachelor’s degree at the Massachusetts Institute of Technology (MIT), double majoring in Electrical Engineering & Computer Science and Mathematics. He also received his Master’s degree at MIT, completing his thesis as part of the Natural Language Processing group at the MIT Computer Science & Artificial Intelligence Laboratory. You can find him on Twitter via @youyanggu.
Joined by guest Vivek Farias (MIT).
4:15 About Youyang Gu
5:09 COVID-19 ML Method
44:10 Advantages and Analysis of the ML Approach