Analysis of Covid-19 Disease Using Machine Learning - Personalized Model

Abstract

he use of artificial intelligence,especially machine learning methods in creating models that will be applied in clinical practice has reached its peak with the appearance of the COVID-19 pandemic. This study aims to determine the severity of the clinical condition of COVID-19 patients based on blood marker analysis. The study used data from 60 COVID-19 patients treated at the Clinical Center Kragujevac. The research methodology includes the selection of the most important laboratory parameters as well as the classification of patients depending on them using methods of supervised learning,regression and classification. With an accuracy of 90%,three parameters were selected that can mostly indicate the severity of the patient's condition,which are: lactate dehydrogenase (LDH),C-reactive protein (CRP),white blood cells (WBC). Laboratory biomarkers such as LDH,CRP and WBC may have an impact on predicting outcomes and help classify patients into an appropriate group based on symptoms.