Cardio-Vascular Disease Classification Using Stacked Segmentation Model and Convolutional Neural Networks

Abstract

Various decision support system based on Artificial Neural Networks have been extensively used for predicting cardio-vascular disease. But, some of the investigations concentrate on preprocessing the features. Here, this work focuses on feature refinement with segmentation and removal of problems related to prediction model. The problems are related to over-fitting and under-fitting. By avoiding these problems, the proposed model shows superior functionality while considering the available dataset. For eliminating the un-necessary parts of input data, denoising stacked encoder is used for configuring the Convolutional Neural Networks with testing and training data. The anticipated model is compared with existing approaches and reported that this model outperforms the existing approaches for predicting heart disease. The anticipated model acquires finest prediction accuracy. The results are seems to be more promising while compared to the other. The findings based on this study recommend that this diagnostic system is utilized for physicians to predict heart disease accurately. Simulation is done in MATLAB environment.