Heart Diseases Diagnosis based on a Novel Convolution Neural Network and Gate Recurrent Unit Technique

Abstract

Actually, one of the leading causes of death is cardiac diseases so medical diagnosis tries to recommend the most candidate diagnose any kind of cardiac disease. Researchers have several distinctive hybrid techniques by strengthening a variety of machine learning methods that aid specialists in the field of cardiac disease expectations. This paper presented a technique named “Convolution Neural Network and Gate Recurrent Unit (CNN GRU).” The main goal of this methodology is to suggest an optimal machine learning approach that achieves high accuracy in the prediction of cardiac disease. The Linear Discriminant Analysis (LDA) and Principal Component Analysis (PCA) feature selection algorithms are utilized to extract essential features from the data set. The proposed technique was compared to several machine learning algorithms with the selected features. The “K-fold” cross-validation was utilized to enhance the accuracy. The results showed that the (CNN GRU) technique achieved 94.5 percent accuracy compared to other techniques.