PREDICTION OF DISEASE STAGE BY MACHINE LEARNING CLASSIFICATION METHODS FOR COVID-19 PATIENTS

CUKUROVA 8th INTERNATIONAL SCIENTIFIC RESEARCHES CONFERENCE 15 - 17 April 2022 / Adana, TURKEY, Adana, Türkiye, 15 - 17 Nisan 2022, ss.1-10

ABSTRACT

In recent years, classification methods, one of the main applications of machine learning, are

widely used in many fields. Among these areas, health is an important area where machine

learning studies are applied successfully. In this study, it is aimed to develop models that

predict disease stage in people with Covid-19 diagnosis using machine learning (ML)

classification methods. Covid-19 is an epidemic disease that was declared a pandemic by the

World Health Organization in March 2020 and caused the death of millions of people around

the world. Today, millions of people still suffer from this disease and face death. In addition

to the problems of medical system inadequacies such as lack of beds, intensive care

occupancy, and respiratory (ventilator) device shortages, the pandemic has also left healthcare

workers faced with the overwhelming burden of patients. For this reason, the ability to detect

the deterioration of the patient by early determination of the disease stage during their stay in

the hospital for Covid-19 patients is very important for hospital management. Within the

scope of the study, clinical and laboratory data of Covid-19 patients at hospital admission

were used. For the data set, models that provide prediction of disease stage were obtained by

using the classification-based machine learning algorithms Logistic Regression, Random

Forest and Support Vector Machines. With the models obtained, the hospital management

will be informed about the number of beds that will be required for moderate, severe and

critical Covid-19 patients and the need for detailed human resource power to take various

precautions in advance.

Keywords: Covid-19, Supervised Learning, Logistic Regression, Random Forests, Support

Vector Machines