Predicting COVID-19 Mortality Rates: An Analysis of Case Incidence, Mask Usage, and Machine Learning Approaches in U.S. Counties

ABSTRACT: The COVID-19 pandemic has necessitated the use of multidisciplinary approach to assess public health interventions. Data science has been widely utilized to promote interdisciplinary collaboration especially during the post-COVID era. This study uses a comprehensive dataset, including mask usage and epidemiological metrics from U.S. counties, to explore the correlation between public compliance with mask-wearing guidelines and COVID-19 mortality rates. After employing machine learning approaches such as linear regression, decision tree regression, and random forest regression, our analysis identified the random forest model as the most accurate model in predicting mortality rates due to its efficacy with the lowest error metrics. The models' performances were rigorously evaluated through error metric comparisons, highlighting the random forest model's robustness in handling complex interactions between variables. These findings provide actionable insights for public health strategists and policy makers, suggesting that enhanced mask compliance could significantly mitigate mortality rates during the ongoing pandemic and future health crises. KEYWORDS: machine learning applications, predictive modeling for public health, COVID-19 analysis, pandemic, model comparison.