PRINCIPAL COMPONENT ANALYSIS AND HYBRID FUZZY CONVOLUTION NEURAL NETWORK TECHNIQUES BASED HEART DISEASES PREDICTION MODEL

Abstract

Heart diseases cause greater mortalities globally which can be reduced by detecting them in their early stages where Clinical Decision Support Systems (CDSS) have been designed. Healthcare organizations use Hybrid Convolutional Neural Network (HCNN) algorithms to determine a patient's risk of dying or being hospitalized from heart failure. However, because of a max-pool operation, a convolutional neural network operates much more slowly. If the computer lacks a powerful GPU, CNN's consume longer times during training as they have several layers. This work presents an efficient model for predictions of cardiac illnesses. Preprocessing is initially carried out using min-max normalization. Principal Component Analysis (PCA) is used in Dimensionality Reduction. Hybrid Fuzzy Convolution Neural Network (HFCNN), based on optimized modified chicken swarm, is used to identify cardiac disorders. This work’s experimental outcomes demonstrate the suggested model’s higher degrees of accuracy in identifying cardiac illnesses.