PREDICTION OF MORTALITY RATE FOR COVID-19 PATIENTS

The aim of this research project is to study and analyze the factors affecting the criticality of COVID-19 patients, and accurately predict the mortality rate of the patients ahead of time. In this paper, COVID-19 data from the National Center for Data of Health which consists of data from 2019 to 2022. Different visualization techniques were used to extract patterns from the demographic and the clinical data of patients to determine the factors affecting COVID-19 patients. Random Forest and Autoencoder neural networks were trained to predict the mortality rate of the patients. Predictions were evaluated using AUC, ROC and accuracy scores. Neural Network resulted in an accuracy of 71.10% and Random Forest gave an accuracy of around 53%.