Hybrid optimization enabled deep learning-based ensemble classification for heart disease detection

Heart diseases (HD) in humans are the most common cause of death. In the current global environment, the early detection of HD is a challenging process. The goal of this work is to develop a deep learning technique and to test the necessary classification model to improve HD detection. Hybrid optimization deep learning-based ensemble classification for heart disease is devised in this research for HD detection. Here, the input data are acquired from the dataset and preprocessed. Then, preprocessed data are subjected to the feature fusion scheme that is carried out by congruence coefficient and overlap coefficient enabled deep belief network. Consequently, with the feature fusion output, heart disease prediction classification is done by the proposed social water cycle driving training optimization (SWCDTO) ensemble classifier, which is devised using the driver training-based optimization algorithm and social water cycle algorithm. This method can efficiently train multiple classifiers to improve their efficiency. These results are combined to produce the final results. Moreover, the introduced SWCDTO-based ensemble classifier approach compared with different heart disease prediction algorithms shows better performance regarding the evaluation measures such as specificity, accuracy, and sensitivity with better values of 95.84%, 94.80%, and 95.36%. Overall the proposed method has low computational time and thus improves efficiency.