Abstract
This study explores how well Artificial Neural Networks (ANNs) can predict the spread of COVID-19 across Canadian health regions, focusing on the impact of socio-economic factors. By examining a wide range of demographic, economic, and social indicators, we identify which factors play the biggest role in accurately forecasting the pandemic’s spread. The trained ANN model underscores the critical role of urbanization, population density, and social behaviors in densely populated regions, such as Toronto and Montreal, where transmission rates were higher. Conversely, remote regions like the Keewatin Yatthé and Labrador-Grenfell Health Authorities saw lower transmission due to geographic isolation and community-based controls. Additionally, the study highlights disparities in healthcare infrastructure, especially in ICU bed availability, which were more pronounced in urban areas. Vaccination rates were also identified as key in controlling the spread, with proactive public health efforts leading to higher rates in regions like the Northwest Territories. Our findings show that these socio-economic factors vary in importance from one region to another, offering valuable insights for public health planning. These findings provide practical advice for improving how resources are allocated and how public health strategies are developed, emphasizing the need to consider socio-economic differences in pandemic forecasting. This approach aims to help policymakers and health officials respond more effectively to current and future public health challenges.