<![CDATA[This study implements a Rough Neural Network (RNN) intelligent system method, merging rough sets with neural networks to diagnose heart disease effectively. Using the Cleveland Heart Disease Dataset, rough sets identified nine relevant features for model training, simplifying data complexity. Comparative assessment against traditional neural networks revealed the RNN model's superior performance, achieving 88.52% accuracy, 88.14% F1 score, and 88.85% AUC. This hybrid approach improves predictive accuracy while enhancing efficiency and interpretability. The findings contribute to advancing intelligent systems for heart disease diagnosis, facilitating early detection, and improving patient outcomes. Future research may explore selected features' clinical significance and RNN applicability in different contexts. Keywords: Heart Disease Detection, Rough Neural Network, Rough Set Theory, Neural Networks, Hybrid Intelligent System]]>
Copyrights © 2024
