HEART DISEASE PREDICTION BY INTEGRATING THE FEATURE OF FUSION WITH MARINE PREDATORS’ ALGORITHM AND ENSEMBLE CLASSIFIERS

Abstract

Heart disease (HD) is a severe and life-threatening condition that affects approximately one-third of global deaths, with an annual toll of around 17.9 million lives lost worldwide. Nearly half of all individuals diagnosed with heart disease succumb to the condition within just 1-2 years, underscoring its significant impact on human health. An estimated 3% of the total healthcare budget is allocated towards treating heart disease. Predicting heart disease requires multiple tests, and inaccuracies may arise due to lack of expertise among medical personnel. Early diagnosis presents challenges, particularly in developing nations where a scarcity of trained medical professionals and essential diagnostic equipment impedes proper patient care. An accurate assessment of cardiac failure risk holds immense potential for preventing severe heart attacks and enhancing patient safety. This study proposes an ensemble classifier for heart disease prediction using three distinct algorithms: support vector machine (SVM), random forest, and naïve Bayes. The Marine Predators Algorithm (MPA) is used to enhance feature selection and optimization, resulting in an impressive accuracy of 97.05% using just five features. The primary objective of this study is to advance upon existing methodologies by introducing an innovative approach to model construction and develop a model that is not only effective but also easily implementable in practical settings, ultimately contributing to improved heart disease prediction and management.