A Hybrid CNN-Transformer Model for Enhanced Heart Disease Prediction

Heart disease is one of the leading causes of death worldwide, making early and accurate prediction essential for timely treatment. Traditional machine learning models like Support Vector Machines (SVM) and XGBoost have been widely used but struggle to capture complex relationships in patient data. In this study, we propose a hybrid CNN-Transformer model that combines the feature extraction power of Convolutional Neural Networks (CNNs) with the long-range dependency learning of Transformers. Our model was tested on the Cleveland Heart Disease dataset and achieved an accuracy of 97.3%, outperforming existing methods like SVM (90.05%) and XGBoost (85.25%). It also showed high precision (97.1%), recall (98.53%), and F1-score (98.97%), demonstrating its reliability in heart disease prediction. These results suggest that combining CNNs with Transformers can significantly enhance predictive performance, offering a more effective and intelligent approach for clinical diagnosis.