Classification of influenza H1N1 and COVID-19 patient data using machine learning

Abstract

COVID-19 is a community-acquired infection with symptoms resembling those of influenza and bacterial pneumonia. It has negatively affected the entire world in areas such as the economy, social life, education, and technology. COVID-19 and H1N1 influenza have been compared in recent studies as they are both causative agents of pandemics and have both caused great distress around the world. Since these two diseases have some symptoms and diagnostic features in common, it would be beneficial for health professionals and scientists to analyze and study patient's clinical data for these two diseases. In this work, we propose some machine learning algorithms to classify patient data into the two classes of H1N1 and COVID-19. The study includes 1467 patient data (70% from H1N1 and 30% from COVID-19) with 42 attributes used in classification. Experimental results show that the Bayes network gives 86.57% accuracy, the naive Bayes classifier gives 82.34% accuracy, the multilayer perception algorithm gives 99.31% accuracy, the locally-weighted learning algorithm gives 88.89% accuracy, and random forest gives 83.16% accuracy for the same data set.