Heart Disease Detection: A Comprehensive Analysis of Machine Learning, Ensemble Learning, and Deep Learning Algorithms

Heart disease is a major health concern impacting a significant number of individuals. Early detection is crucial for effective treatment and management. Conventional techniques for the detection of heart diseases are time-consuming, inconvenient, expensive, and unsuitable for diagnosis. To take advantage of innovative approaches such as artificial intelligence (AI), this paper presents the detection of heart diseases at an early stage and enhances the accuracy of AI models for predictions, because accurate identification is crucial in the early stages, if not detected promptly, it may lead to death. Two methods are employed for the early detection of heart disease. In the first approach, traditional machine learning, ensemble learning, and artificial neural networks are utilized. In the second approach, a hybridization approach is applied to machine and ensemble learning algorithms to boost model performance. The heart_statlog_cleveland_hungary_final dataset is utilized and split using k-fold cross-validation with a value of k set at 10. The performance metrics such as accuracy, sensitivity, specificity, precision, and F1 score are calculated. The results indicate that the hybrid technique, namely Bagging combined with random forest (RF), emerges as the top performer, boasting the highest average accuracy of 94.34%, average specificity of 93.7%, average sensitivity of 93.5%, average precision 94%, and an average F1 score of 94.2%. In conclusion, the hybrid approach of Bagging with RF would be better for detecting heart disease at an early stage.