LASSO-DT Based Classification Technique for Discovery of COVID-19 Disease Using Chest X-Ray Images

Early recognition of COVID-19 can aid in the development of a treatment plan and disease containment decisions. This research aims to use the LASSO+DT technique to automatically diagnose COVID19 lung-related individuals employing automated chest X-ray imageries while optimizing detection accuracy. The study employed machine learning (ML) based techniques, specifically feature selection (FS) and classification models were used to classify COVID-19 and normal chest X-ray imageries. For the FS, Least Absolute Shrinkage and Selection Operator (LASSO) technique was employed and for the classification, a Decision Tree (DT) classifier was used. A dataset containing 372 instances of chest X-rays was employed in this research for the investigation. The classification accuracy, detection rate (DR), and false-positive rate (FPR) were employed for the performance evaluation of the investigation. The study implemented DT alone as well as LASSO+DT and the results of both classifiers were compared. The result of the study recognized that the proposed LASSO+DT outperformed that of the DT alone. The investigation demonstrates that the proposed LASSO+DT is efficient and effective for the identification of COVID-19 ailment and further lung diseases because it has higher accuracy, DR, and lower FAR when compared with DT technique and few state of the art.