Early diagnosis of heart disease using machine learning and explainable artificial intelligence: a robust framework for improved healthcare outcomes

Painuli D., Mishra D., KÖSE U., Pragya P., Shaban M.

Explainable AI in Clinical Practice: Methods, Applications, and Implementation, Elsevier, ss.327-342, 2026

  • Yayın Türü: Kitapta Bölüm / Araştırma Kitabı
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1016/b978-0-443-44111-0.00019-6
  • Yayınevi: Elsevier
  • Sayfa Sayıları: ss.327-342
  • Anahtar Kelimeler: disease diagnosis, explainable AI, feature engineering, heart disease, hyperparameter tuning, Machine learning, medical diagnosis
  • Süleyman Demirel Üniversitesi Adresli: Evet

Özet

Early diagnosis of health conditions, particularly heart disease, is critical for improving patient outcomes and reducing mortality rates. This study proposes a machine learning (ML)-based methodology for the early diagnosis of heart disease using symptom-based datasets. The framework integrates explainable artificial intelligence (XAI) to enhance the interpretability and trustworthiness of the model’s predictions, addressing the need for transparency in healthcare applications. To address class imbalance in the dataset, we employ the Synthetic Minority Over-sampling Technique, ensuring balanced representation of positive (disease) and negative (nondisease) cases during model training. Three state-of-the-art ML algorithms—support vector machine (SVM), random forest (RF), and Extreme Gradient Boosting (XGBoost)—are evaluated using fivefold cross-validation and GridSearchCV for hyperparameter optimization. Experimental results demonstrate that the XGBoost-based model achieves superior diagnostic accuracy (97.78%) compared to SVM and RF. The model’s decision-making process is further validated using SHAP (SHapley Additive exPlanations), an XAI technique, to provide interpretable insights into feature contributions. This study presents a robust, transparent, and effective framework for the early detection of heart disease, leveraging advanced ML techniques and symptom-based data to enable timely interventions and improved healthcare outcomes.