Hybrid technique for lung disease classification based on machine learning and optimization using X-ray images

The remarkable growth of machine learning has shown that it can perform at an expert level in a number of difficult tasks, such as medical decision-making and image processing. The goal of this research is to take advantage of this potential by creating a new and precise technique for chest X-ray analysis-based lung disease identification. The goal of this work is to research the methods currently used for creating the architecture of machine learning and create our own model, which will then be used for the best possible detection of lung illness using digital medical (X-ray) pictures. Therefore, the focus of this work is on employing various machine learning approaches to diagnose lung illness. The proposed methodology's optimization section consists of 2 primary phases. In the first phase, bagging, voting ensemble learning and boosting techniques are utilized to enhance performance of the support vector machine (SVM), K-nearest neighbours (KNN), and naïve Bayes (NB) algorithms. Emperor Penguin Optimization (EPO) was then used to the system to optimise the parameters and improve the overall system performance. The outcome shows the efficacy of the proposed methodology that compared to other hybrid techniques, the combination of ensemble learning and EPO with SVM produced the highest accuracy of 97.50%. This result suggests that the proposed system has the potential to greatly increase the precision of conventional machine learning methods in the identification of lung illness using chest X-rays. This development may lead to more accurate diagnoses and better patient outcomes.