EDL-NSGA-II: Ensemble deep learning framework with NSGA-II feature selection for heart disease prediction

With the onset of pandemics and accrescent stress among people, heart disease has become one of the leading contributors to premature deaths worldwide. Early prognosis of this disorder is the most viable strategy for increasing the prospects of individuals' survival. Numerous methods exploiting machine learning algorithms for heart disease prediction have been reported in the literature, but they all suffer from overfitting problems. Conspicuously, to improve the prediction accuracy, in this research work, an efficient meta-heuristic-based feature selection technique, namely NSGA-II, is employed. The proposed solution aims to reduce the feature set and thus improve the prediction accuracy supported by intelligent machine learning models. The presented classifier is trained and validated using a comprehensive heart disease dataset by utilizing NSGA-II selected features with 10-fold cross-validation. For performance validation, the results of the proposed model are compared with the state-of-the-art machine learning algorithms, such as a k-nearest neighbour, support vector machine, Bayesian belief networks, random forest, and naive bayes. The simulation results highlight the improved performance of the proposed model with NSGA-II and achieve a high prediction accuracy of 97.32%. Furthermore, results of sensitivity (92.84%), specificity (92.60%), precision (91.25%), and F-measure (92.17%) prove the utility of the proposed approach for heart disease prediction over all other variants.

Automated summary

With the increase in pandemics and stress levels, heart disease has become a leading cause of premature deaths worldwide. This research work proposes the use of the NSGA-II feature selection technique to improve prediction accuracy and achieve a high prediction accuracy of 97.32% for heart disease. The results demonstrate the utility of this approach over other variants.