Performance of Machine Learning Models for Pandemic Detection Using COVID-19 Dataset
Sunday Adeola Ajagbe, Adekanmi Adeyinka Adegun, Pragasen Mudali, Matthew O. Adigun
Abstract
The pandemic produced by coronavirus2 (COVID-19) and other related infectious diseases have been confined to the world, and there is a need to control its spread as well as prepare for any related outbreak although early detection strategies. Therefore, this paper aimed to identify an efficient machine learning (ML)-based model for pandemic detection to combat the spread of pandemic and related infectious diseases. Seven (7) ML-based models are studied: k-nearest neighbor (KNN), support vector machine_poly, (SVM-Poly), support vector machine_RBF, random forest (RF), decision tree (DT), XGBoost, and Logistic regression (LR) were used for quick and better detection of potential COVID-19 cases. The dataset utilized picks the pertinent symptoms for the identification of a suspicious person from COVID-19 symptoms. The experiments achieved the XGBoost leading with an accuracy of 98.4%, a precision of 94.0%, a recall of 93.5%, and F1-Score of 94.0% respectively. The results showed that real-time data capturing will efficiently detect and monitor COVID-19 patients. This research will help many research teams create useful apps based on ML, DL, and AI models, as well as help healthcare organizations, academics, and governments by demonstrating how these methods can make it easier to combat COVID-19.