| X-Heart: A Hybrid Explainable AI Framework Integrating Genetic Algorithms and Machine Learning for Heart Disease Diagnosis |
| Paper ID : 1069-ICEEM2025 (R2) |
| Authors |
|
Noha Fathalla Mahmoud *1, Ezz El-Din Badawi Gad El-Rab2, Prof. Ayman El-Sayed Omera2 1Computer Engineering, Electronic Engineering,Menoufia University,Menouf,Egypt 2Computer Science and Engineering Department, Faculty of Electronic Engineering, Menoufia University, Menoufia, Egypt |
| Abstract |
| Heart disease continues to be the leading cause of death globally, responsible for an estimated 17.9 million deaths annually. The importance of early and accurate diagnosis cannot be overstated, as it enables timely clinical intervention that can significantly improve patient outcomes. Recent advancements in artificial intelligence (AI) have created new opportunities to enhance diagnostic methods. This study proposes a novel hybrid diagnostic framework that integrates Genetic Algorithms (GA) with Machine Learning (ML) models to improve the detection and analysis of heart disease. The framework incorporates ensemble classification techniques under the umbrella of Explainable Artificial Intelligence (XAI), aiming to balance predictive performance with model transparency and interpretability. By combining evolutionary optimization with XAI principles, the proposed GA-ML framework offers a powerful and explainable approach for analyzing cardiovascular data. The framework was rigorously evaluated using two benchmark datasets: Statlog and Cleveland. Experimental results show that the hybrid model consistently outperforms traditional ML models across multiple evaluation metrics. For example, on the Statlog dataset, the GA-ML model achieved accuracies of up to 97% with Random Forest, compared to a maximum of 90% for standalone ML models. Similarly, for the Cleveland dataset, the hybrid model attained superior performance across most classifiers. These findings demonstrate the framework's effectiveness in enhancing diagnostic precision, reliability, and transparency. Ultimately, this research provides a valuable tool for healthcare professionals, contributing to more informed decision-making, reducing diagnostic uncertainty, and supporting efforts to lower global heart disease mortality rates through improved technology-driven diagnosis. |
| Keywords |
| Heart Disease Diagnosis, Artificial Intelligence, Machine Learning, Genetic Algorithms, Hybrid Models, Random Forest, K-Nearest Neighbors, Logistic Regression, Medical Decision Support |
| Status: Accepted |