Decision Tree for Determining Hospital Treatment for Covid-19 Patients Based on Hematology Parameters Using the C5.0 Algorithm

Abstract

The rapid spread of the COVID-19 disease which occurred globally throughout the world some time ago, requires early detection of COVID-19 which is very important for patients and also the people around them to be able to fight the COVID-19 pandemic. Therefore, a classification analysis will be carried out to make decisions regarding determining COVID-19 patients who do not require hospitalization or who require Regular Ward, Semi-Intensive Care Unit, or Intensive Care Unit (ICU) in hospitals based on hematology parameters from the Machine Learning Repository. Kaggle Dataset uses the C5.0 algorithm assisted by Rstudio software. It is also known that because the data contains missing data, it is also necessary to handle missing data using the Mean Method assisted by SPSS software. Performance evaluated using the Confusion Matrix method produces an accuracy value of 78% which is considered quite good, where testing with the C5.0 Algorithm uses a training and testing data ratio of 40:60. This research was carried out with the aim of simplifying and speeding up the performance of medical personnel so that COVID-19 patients receive fast and appropriate treatment to help reduce COVID-19 cases in a population.

Keywords

• early detection
• classification
• missing data
• confusion matrix