An Artificial Neural Fuzzy Inference System Based on a Grey Wolf Optimized Convolution Neural Network for Efficient Cardiovascular Disease Prediction

Abstract

Abstract: Increasing variations in daily human lifestyle, such as food habits, work routines, and modern communication methods, have significantly influenced human health. Cardiovascular diseases (CVDs) remain among the most prevalent and life-threatening conditions, often leading to sudden cardiac arrest, and other chronic diseases. The advancement of big data and Artificial Intelligence (AI) has revolutionized disease detection, yet challenges persist due to the complex nature of disease properties and unknown feature dependencies. To resolve these problems, an ANFIS-based CVD disease prediction system using a Grey Wolf Optimized Convolutional Neural Network (GWO-CNN) is proposed in the paper. Heart Rate Variability (HRV) is utilized to estimate disease impact rates, followed by Structural Cascaded Pattern Mining (SCPM) to construct disease patterns based on high-impact margins. Feature selection is conducted using the ANFIS model, which effectively reduces irrelevant features, and the classification is performed using GWO-CNN to identify CVD risk levels. The experimental results section evaluates the model’s performance through accuracy, precision, recall, specificity, and F1-score, demonstrating superior results compared to conventional methods. The FGWO-CNN model, which integrates GWO with a Deep CNN (DCNN), achieved superior performance compared to baseline models such as Random Forest with an Improved Linear Model (RERF-ILM) and Queen McCluskey Binary Classifier (QMBC). The proposed system achieves high performance, improving prediction accuracy up to 96% while maintaining a low false positive rate and reduced time complexity. This approach enhances early CVD detection, enabling timely interventions and improving health outcomes through performance measurement.