A Multi-Folded Dynamic Regularized Dual Crossed CNN with A Self-Adaptive Metaheuristic Aware Cardiovascular Disease Prediction

Abstract

Cardiovascular Disease (CVD) prediction is becoming a complex task in research, where enormous work is being carried out to provide a highly efficient prediction system. Deep learning-based prediction is a highly significant task for achieving a robust detecting system, which prevents death through early diagnosis. However, the existing system has the challenges of overfitting, improper feature learning, imbalanced dataset, high false positive rate, and high false negative rate. Therefore, the proposed model has introduced a novel multi-folded regularized Convolutional Neural Network (CNN) based on dynamic regularized dual crossed CNN, Fuzzy C-Means based feature aggregation, a Self-adaptive Whale Optimization Algorithm (SWOA) with the Optimized Information Gain (IG). The proposed multi-folded dynamic regularized dual crossed CNN framework, implemented in Python, achieves superior performance in heart disease prediction, with metrics such as accuracy (99.1234), Sensitivity (98.9876), specificity (99.4321), and also other metrics significantly outperforming existing methods in terms of Precision, F-Measure, MCC, NPV, FPR, and FNR.