Chronic SARS-CoV-2 (COVID-19), Lung Cancer, Tuberculosis, and Pneumonia are among the major worldwide health issues. Early detection and prompt medical treatment are crucial to enhancing patient survival and minimizing death rates, particularly in instances with extreme respiratory diseases. Although deep learning (DL) and machine learning (ML) models have shown immense success in classification tasks on images, dependence on large databases could be a drawback in this model too, especially in the event of global health emergencies like the COVID-19 pandemic. In this article, we propose a bio-inspired approach in lung disease classification by integrating two algorithms: Social Group Optimization (SGO) and Particle Swarm Optimization (PSO). The introduced model effectively selects informative features from chest X-ray (CXR) images so that the classification can be made accurately through ML methods like Support Vector Machines (SVMs), Naive Bayes, Linear Discriminant Analysis (LDA), and Random Forests. Experimental findings reveal that the pipeline registers 97.1% classification accuracy using SVM, and hence, it could prove to be an efficient tool in early diagnosis of lung diseases.