RESILIENCE GREY WOLF OPTIMIZATION-BASED CONVOLUTIONAL NEURAL NETWORK (RGWO-CNN) FOR CLASSIFICATION OF HEART DISEASE AND DIABETES (HDD)

Heart Disease and Diabetes (HDD) has reached epidemic proportions worldwide. Long-term issues affecting human organs are more likely due to HDD. If this critical medical condition is diagnosed at an early stage, people may be able to live healthier and longer lives by taking proper medications. The properly trained machine learning models on relevant datasets will help to diagnose HDD, but the significant issue is the lack of classification accuracy. Even more importantly, most current machine learning algorithms focus on predicting specific diseases. A classifier that accurately predicts the incidence of several diseases might be helpful in this context. This paper proposes a Resilience Grey Wolf Optimization-based Convolutional Neural Network (RGWO-CNN) to classify HDD. In the proposed RGWO-CNN method, the hyperparameters of the CNN model are represented as individual wolves in the RGWO algorithm. The position of each wolf represents a specific set of hyperparameters. The RGWO algorithm iteratively updates the position of each wolf based on their fitness (performance) and the wolves' social hierarchy (dominance). The updated position of each wolf corresponds to a new set of hyperparameters used to train and evaluate a new CNN model. This work uses RGWO-CNN to the Cardiovascular Disease Dataset and the PIMA Indian Diabetes Database to evaluate its performance. The assessment results reveal that the proposed classifier offers superior classification accuracy compared to the state-of-the-art classifiers.