Artificial Intelligence COVID-19 School Testing Effectiveness Application



Advanced Disaster, Emergency and Rapid-response Simulation (ADERSIM)

York University


DESCRIPTION

This application helps you to estimate potential impacts of COVID-19 outbreaks in a school and assess the effectiveness of different COVID-19 testing strategies.

APPROACH

The Artificial Intelligence COVID-19 School Testing Effectiveness Application has been developed based on a big dataset extracted from more than 125K runs of a school testing simulation (https://cloud.anylogic.com/model/a7c4411e-064e-4283-a93c-b0b27e0430ee). The simulation uses an agent-bases model incorporating an adjusted version of the SEIR (Susceptible, Exposed, Infectious, Recovery) to account for testing and self isolation. A preprint of the article with LANCET is available (10.2139/ssrn.3699573). The AI model is a basic fully connected feed-forward network. Our AI model has a mean absolute error of 14.6 on the test data (measured over the entire output vector), or more intuitively an average absolute error of just 0.01098 to each individual output data point.

RESEARCH TEAM



DISCLAIMER

The Information is licensed as is, and the Information Provider excludes all representations, warranties, obligations, and liabilities, whether express or implied, to the maximum extent permitted by law.

ACKNOWLEDGEMENT

The team acknowledge the support received from a number of funding agencies including: Public Health Agency of Canada (PHAC), Canadian Institute for health research (CIHR), and Ontario Research Fund.