Modeling the COVID-19 pandemic
using machine learning to predict disease outcomes
At the start of the pandemic, the members of the Caltech community enrolled in the course CS-156 came together to build predictive models of the rapidly spreading disease, COVID-19. Different teams contributed a variety of models to predict COVID-19 outcomes with uncertainty quantification on a per-county basis over a two week period. The final predictions, which consisted of an ensemble of the best performing team models, included epidemiological, deep learning, and regression models. We then collaborated with the CDC-funded UMass-Amherst Influenza Forecasting Center of Excellence, by contributing our predictions to their forecast hub. The goal of this hub is to compile, standardize, and ensemble predictions from groups across the country to be used for policy decisions and visualization.
Related Links:
https://www.caltech.edu/about/news/students-use-ai-better-prediction-covid-19-model https://www.caltech.edu/about/news/caltechs-ai-driven-covid-19-model-routinely-outperforms-competitors https://magazine.caltech.edu/post/covid-19-research-update