Leveraging IBM Watson Auto AI to Predict the Occurrence of Deaths Among COVID-19 Patients in USA

Abstract

Technology and preventive healthcare have become the key solutions to the COVID-19 pandemic. This study explores the COVID-19 dataset to predict the occurrence of deaths among COVID-19 patients in the USA. 1,386,100 observations and 5 variables were selected from the USA COVID-19 Dataset obtained from the online repository of the Center for Systems Science and Engineering (CSSE) on Johns Hopkins University's (JHU) website. IBM's Auto AI Experiment was leveraged to determine the best prediction outcomes. All eight pipelines were trained and tested using the COVID-19 dataset and the results showed that Pipelines 3, 4, 5, 6, 7, 8 outperformed the other two algorithms in terms of accuracy, precision, recall, and the F scores. Pipelines 1 and 2 achieved the worst results among the models with an F score of 0.9915. The research demonstrates the promising performance to help health institutions and governments plan ahead to forestall COVID-19 deaths and implement policies toward preventive care.