COVID-19 Symptoms Detection Using Iterative Dichotomiser (ID3) Algorithm: A Greedy Method
Hassan Adamu, Azman Samsudin, Jamilu Awwalu
Abstract
Iterative Dichotomiser 3 (ID3) algorithm is a popular greedy algorithm that has been applied in solving different types of classification and prediction problems. With the current COVID-19 pandemic, authorities are pushed to the limit on resource allocation and laboratory test administration for suspected cases. In this study, an ID3 model is created using the holdout and cross-validation approaches to classify potential COVID19 cases based on presented symptoms. The two approaches in the ID3 model are compared based on accuracy and time complexity. Findings from the study show that the 50:50 split holdout approach achieved the overall best accuracy across both the holdout and cross-validation approaches, while the crossvalidation had the highest time complexity.
Index Terms—Greedy algorithm, Iterative Dichotomiser (ID3), COVID-19 symptoms classification