Machine Learning-Driven Cardiovascular Disease Prediction: Balancing Performance and Transparency with Hybrid Models