NORMAN, OKLA. – New research out of the University of Oklahoma provides better insight into predicting how the coronavirus might spread in rural communities.
Led by Charles Nicholson, an associate professor in the School of Industrial and Systems Engineering, Gallogly College of Engineering, the project applies machine learning techniques and data like social media, mobility, demographics and health factors to develop models to predict the spread of COVID-19.
“We wanted to primarily focus on developing models for rural areas and counties for whom the existing forecasting models were limited or inaccurate due to small sample sizes,” Nicholson said. “Our ultimate goal is to provide early warning and actionable information to medical service providers.”
Nicholson said the rapid spread of the pandemic made it difficult to develop accurate models.
“Particularly in the early months of the pandemic, many forecasts were made regarding the potential number of cases and fatalities associated with the virus; however, much of that data skews toward large urban areas,” he said.
One of the reasons more localized modeling is necessary, he added, is to help policymakers make more informed decisions for their communities.
“While large, population-dense areas see more infectious disease cases, they also have greater resources,” Nicholson said. “For example, Oklahoma’s budget per capita is only 52.6% of New
York’s. Accurate disease projections are potentially ‘mission critical’ for Oklahoma given this relative shortfall of resources.”
By creating county-level, community-specific COVID-19 forecast models designed explicitly for small populations, Nicholson said they “…can help policy makers confront the threat of the virus without overextending their limited financial resources.”