CMRE is collaborating with researchers at the University of Connecticut (UConn) to develop a new modelling approach to forecast the spread of COVID-19.

CMRE’s Data Knowledge and Operational Effectiveness (DKOE) team and their UConn colleagues are contributing to the Connecticut Academy of Science and Engineering (CASE) council that is advising the Governor of Connecticut on strategies to "re-open" the State. The CASE council comprises a multidisciplinary team of engineers, mathematicians, social and medical scientists, and government experts.

DKOE’s contribution was to adapt their knowledge of tracking and predicting the movement of targets (such as ships, submarines, and aircrafts) to tracking and forecasting the number of people in a population infected with coronavirus. They applied a technique called Adaptive Bayesian Learning to a number of epidemiological models, which allowed key parameters, such as the infection rate, to be modelled as time-varying instead of stationary. The proposed approach is validated against COVID-19 epidemiological data from Italy and the United States. The DKOE team and their UConn colleagues are about to submit a paper for peer review which demonstrates the performance of their approach for epidemiological modelling.

Read on NATO News: https://www.nato.int/cps/en/natohq/news_177078.htm?utm_source=twitter&utm_medium=NATOPress&utm_campaign=20200703_cmre

Image Caption:
CMRE staff discuss a graph of the reported and modelled Covid-19 infections in the US. Estimated infections during the learning phase of the model are shown as a solid red line. The forecast infections are shown as a dashed line. The vertical dashed line is 26 April 2020, the beginning the forecast. The reported number of infections are shown as a solid blue line. The red area represents the 90% confidence interval of the forecast.