An Explainable AI Model in Heart Disease Classification using Grey Wolf Optimization

Abstract

Heart disease is one of the world’s leading causes of death. It is estimated that around one-third of all deaths are caused by heart disease in the entire world. Recently many research works have focused on using machine learning models to detect and warn patients about the occurrence of heart disease at the early stage. However, machine learning models provide promising results, and the performance of the classification is affected by various reasons which include imbalanced training, and missing values. There are three main contributions of this research work. Firstly, missing values are addressed by employing a grouping of instances. Secondly, a dual filter based feature selection is introduced to pick the most effective features and lastly, we make of Grey Wolf Optimization for optimizing the hyperparameters of the machine learning models. Together, these contributions aim to improve the robustness and efficiency of machine learning applications by addressing missing data, optimizing feature selection, and fine-tuning model parameters. The accuracy of 98.41% indicates the superiority of the proposed classification which is more than 17.15% than the existing machine learning models. On the other hand, we use Explainable AI (XAI) methods to make our proposed model interpretable.