Design and Analysis of a Deep Svm-Based Classifier for Enhanced Lung Infection Detection and Classification

This paper presents a Deep Support Vector Machine (DeepSVM) designed to identify lung infections using a comprehensive set of features. By leveraging this full range of attributes, the model effectively differentiates between infected and non-infected lung samples, contributing to a more precise understanding and definition of lung infections. The data used in this study originates from diverse sources and is strategically divided to optimize both training and testing phases. Approximately half of the dataset is allocated for training the DeepSVM classifier, while the remaining portion serves testing requirements. The training and testing processes are guided by a five-fold cross-validation approach, employing an 80:20 split, which enhances the reliability of the performance evaluation. To assess the efficacy of the DeepSVM model, a series of simulations were conducted. These simulations serve as a critical method to gauge the model's accuracy and robustness, indicating the degree to which the DeepSVM approach excels in classification tasks. Current findings demonstrate that the DeepSVM-based classifier achieves higher classification accuracy than alternative machine learning classifiers available. This notable performance is attributed to the model's capacity to effectively capture complex patterns within the data, leading to enhanced precision in detecting lung infections. The results underscore the DeepSVM model's potential as a leading tool in medical diagnostics, offering superior classification capabilities in comparison to other machine learning techniques commonly used in this domain.