Binary Classification of Lung Diseases Using Convolutional Neural Networks and Support Vector Machines

This study explores the use of a hybrid machine learning model that combines support vector machines (SVMs) and convolutional neural networks (CNNs) for the diagnosis of lung disorders in humans, with a focus on differentiating between healthy and infected lung states. A balanced dataset was used to thoroughly assess the model's performance, and important metrics including precision, recall, F1-score, accuracy, and support were measured. The model obtained an F1-Score of 93.40%, a Precision of 93.17%, and a Recall of 93.63% for the Healthy class. Comparably, the F1-Score was 93.32%, the Precision was 93.55%, and the Recall was 93.08% for the Infected Lung class. A well-balanced assessment was ensured by the fact that each class made up 50% of the sample, with 8010 examples for Healthy and 7950 for Infected Lungs. The model performed well, with a total accuracy of 93%. Additionally, the precision, recall, and F1-score macro, weighted, as well as micro averages were consistently 93.36%, demonstrating consistent efficacy across many assessment criteria. These findings demonstrate the possibility of merging CNNs with SVMs in the analysis of medical images, offering a trustworthy and effective diagnostic tool for the identification of lung diseases. To further enhance and validate the model's performance, the study highlights the significance of advanced machine learning techniques in improving diagnostic accuracy and suggests future research directions, such as dataset expansion, integration of additional imaging modalities, and clinical validation.