acm-header
Sign In

Communications of the ACM

ACM TechNews

Researchers Fix Pandemic Prediction Model, Improving Its Accuracy


View as: Print Mobile App Share:
coronavirus crystal ball, illustration

A white noise model systematically underestimated the severity of the pandemic because of unrealistic noise-induced transitions.

Credit: Getty Images

North Carolina State University's Mohammad Farazmand and Konstantinos Mamis corrected a flaw in a popular pandemic model, improving the accuracy of its predictions.

The researchers tweaked a key parameter in a stochastic compartmental model which segmented populations into different compartments, based on disease status. Each compartment has an associated equation containing certain parameters; health officials predict disease spread by filling in the values for each parameter within the equations.

The model must incorporate a certain amount of noise to mirror the fluctuating contact rate parameter, but Farazmand and Mamis found the addition of random "white noise" caused the compartmental model of the COVID-19 Omicron variant to underpredict disease spread. They applied the Ornstein-Uhlenbeck process to the model to more accurately correlate the forecast to actual data.

From North Carolina State University
View Full Article

 

Abstracts Copyright © 2023 SmithBucklin, Washington, DC, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account