COVID-19 Detection Using Decision Tree Support Vector Machine Random Forest.

Abstract

The rapidly spreading novel coronavirus (SARS-CoV-2) presents a substantial global health risk, necessitating swift detection through efficient diagnostic methods. Reports highlight lung-related issues in many COVID-19 patients, and both chest CT scans and X-rays are effective in identification. Chest X-ray (CXR) imaging, due to its convenience, faster imaging, simplicity, and lower cost, stands out. In our study, we applied Decision Tree, Random Forest, and Support Vector Machine to identify COVID-19 cases from CXR images, comparing their performance to select the most accurate one. Our model achieved a remarkable 95% accuracy in COVID-19 detection and, after further training, attained 87% accuracy in classifying images into Normal, Viral pneumonia, and COVID-19 categories. This model is expected to play a pivotal role in rapidly identifying COVID-19 cases, contributing to a more efficient screening process and disease control. There is a diverse array of machine learning algorithms at our disposal, each providing unique approaches to detect COVID-19 in chest X-rays. This project aims to construct a model that consistently yields the highest level of accuracy when deployed on extensive datasets. This investigation will assess four distinct models, namely Decision Tree, Random Forest, and Support Vector Machine