CORRELATION OF COVID-19 TO LUNG INFECTIONS AND PREDICTION OF LUNG INFECTIONS IN COVID-19 PATIENTS IN IRAQ USING DATA MINING METHODS

Abstract

The Covid-19 pandemic emerged as an unforeseen global crisis, exerting a profound influence on various aspects of human life. Hence, the need for collaborative efforts and scholarly investigations to address and alleviate the challenges arising from this crisis is crucial. One notable concern pertains to lung infections, which are recognized as a highly perilous consequence of the aforementioned virus. Thus, this study aims to investigate the potential correlation between Covid-19 and lung infections, and test the efficacy of various algorithms in predicting lung infections amongst Covid-19 patients. For this purpose, data has been procured from multiple health institutions in Iraq. Using this data, a robust correlation between Covid-19 and lung infection cases was found and the bagging, boosting, naïve Bayes, K-Nearest Neighbour, J48, random forest, PART, and logistic regression algorithms showcased a high accuracy in prediction lung infection in Covid-19 patients, with naïve Bayes achieving the highest accuracy of 93.41 percent.