Effective heart disease prediction using novel MLP-EBMDA approach

Abstract

Heart disease prediction is more important to prevent the death rate. The death rate increases due to lack of initial detection of heart disease in humans. To predict heart disease in an effective way by using feature selection and classification approach. Thus, an optimized unsupervised technique for feature selection and novel Multi-Layer Perceptron for Enhanced Brownian Motion based on Dragonfly Algorithm (MLP-EBMDA) for classification of heart disease has been proposed in this study. The dataset of the heart disease is obtained as input and pre-processing is performed. Features are selected through the optimized unsupervised technique. Based on the selected features, classification of heart disease is performed using the novel hybrid MLP-EBMDA approach. The analysis of the proposed system with various existing systems in terms of accuracy has explored 94.28%. The analysis of the proposed system in terms of precision has showed 96%, in terms of recall the results of the proposed system has been found to be 96%, in terms of F1-score the results of the proposed system has been found to be 96%. Thus the overall performance analysis of the proposed methodology exhibited efficient results in predicting the heart disease than the various state-of-the-art methods.