Ensemble-based Disease Prediction in Medical Image Data

Disease prediction is crucial in the healthcare industry. Early detection and diagnosis of sickness can result of in better patient outcomes and cheaper medical expenses. Machine learning (ML) algorithms have shown considerable potential for sickness prediction jobs due to their ability to learn from vast datasets and find subtle patterns that may not be evident to humans. Multiple machine learning models are used in ensemble learning, a potent method for disease prediction, to increase prediction accuracy and robustness. This paper introduces an ensemble-based method for medical data-driven illness prediction, focusing on chest X-ray datasets for viral pneumonia, lung opacity, normal, and COVID-19. Our goal is to leverage the potential of ensemble methods to increase prediction accuracy and offer practical insights into these diseases. To assess the effectiveness of our ensemble strategy, we have used various machine learning (ML) algorithms such as Support vector machine (SVM), K-nearest Neighbour (K-NN), Naive Bays and Logistic Regression. The ensemble methods used are Random Forest, XgBoost, Decision Tree classifier, stacking and voting. We have compared and analysed the result with various training testing ratio on three well known datasets with five performance measure like AUC, Accuracy, Precision, Recall, and F1-score. As a result of all the models, XgBoost, stacking, voting, and random forest produce the best results.