Optimized convolutional neural network using grasshopper optimization technique for enhanced heart disease prediction

Abstract

According to the World Health Organization (WHO), the heart disease (HD) is a preeminent worldwide cause of mortality. Early prediction and diagnosis of HDs becomes very crucial to save the human kind. This study presents a novel approach by integrating the machine learning (ML) technique, explicitly, a convolutional neural network (CNN) model with grasshopper optimization (GHO) algorithm to optimize the performance of conventional CNN, thereby, the efficiency and accuracy of the proposed HD prediction (HDP) model is enhanced. While evaluating on Cleveland Dataset, the proposed hybridized and optimized CNN model using GHO demonstrated a superior performance metrics, namely, classification accuracy of 88.52%, precision of 87.87%, recall of 90.62% and F1-score of 89.23%. The results emphasize the model’s potential and robustness for early diagnosis, contributing to significant improvements than the conventional ML methods. Further, the proposed study strengthens the growing body of artificial intelligence (AI)-driven healthcare solutions and highlights the significance of hybrid models in the healthcare domain.

Keywords

• Machine learning
• CNN
• optimized CNN
• hyperparameters
• grasshopper optimization
• Artificial Intelligence