An Efficient Framework for Classifying Novel Corona Cases based on Machine Learning and Data Mining Techniques

Abstract

Hospitals with the disease of coronavirus (COVID) are always at risk of dying. COVID hospitalized patients may benefit from the application of machine learning and data mining techniques to predict their death. Therefore, the purpose of this paper was to evaluate many machine learning and data mining algorithms to COVID mortality prediction utilizing patient data at the time of first admission and select the algorithm that performs best as a decision-making tool. Its signs and symptoms resemble those of the regular flu consisting fever, cough, shortness of breath, fatigue, and muscle pain. This paper proposed running the most important algorithms from data mining and machine learning such as naïve bayes, models of decision tree (ID3 and C4.5), support vector machine, and logistic regression to classify corona cases. To test the proposed framework, the confusion matrix has been used. From the confusion matrix the important performance measures have been computed such as accuracy, recall, precision, balance accuracy, and AUC. The experimental findings of this paper supported the notion that the supported vector machine algorithm had good performance and high accuracy in classifying corona disease.