Using Decision Trees to Analyze the Evolution of Covid in Diabetic Patients

Covid-19 has impacted human life in many ways since its worldwide spread. Science, therefore, was faced with a great challenge: to understand the virus that had been causing a staggering number of deaths around the world, as well as to find ways to treat patients and prevent the continued spread of the disease. In this context, it was soon realized that certain groups of people, such as diabetics, have a much higher risk of death when infected with the SARS-CoV-2, virus that causes the disease. Therefore, this paper presents an analysis of the behavior of COVID-19 disease in diabetic patients who have contracted the virus before taking any anti-covid vaccine. The analysis presented in this paper is achieved by employing a robust Decision Tree classifier model to a database containing information about diabetic patients, collected at the University Hospital (HU—UFSCar) of São Carlos, Brazil. Our study leverages machine learning techniques, specifically decision tree analysis and data mining, to unearth patterns and predictors of COVID-19’s impact on diabetic individuals, thereby contributing valuable insights to the field of epidemiology and public health informatics. To assess its quality, the performance of the Decision Tree model is also compared with that of Random Forests. As a result of it, the generated predictive models are useful to the medical staff, both for decision-making about the most appropriate treatment for each individual, as well as for a better understanding of the disease and its evolution in diabetic patients.