کنفرانس بین المللی فناوری اطلاعات و دانش

صفحه اصلی / شانزدهمین کنفرانس بین المللی فناوری اطلاعات و دانش

An Enhanced Fuzzy Rule-Based Method for Coronary Artery Disease Risk Prediction Using Weighted and Biased Rules

نویسندگان :

Fatemeh Ahmadi¹ Mohammad Javad Parseh² Ehsan Amiri³

1- دانشگاه جهرم 2- دانشگاه جهرم 3- مجتمع آموزش عالی لارستان

کلمات کلیدی :

Coronary Artery Disease (CAD)،Fuzzy Inference System (FIS)،Mamdani Inference،Membership Functions،Weighted Fuzzy Rules

چکیده :

The integration of artificial intelligence (AI) and fuzzy inference systems (FIS) has shown strong potential for improving clinical decision support in healthcare. Physicians can benefit from AI-driven analyses applied to electronic health records (EHRs), enabling the identification of shared patterns among patient cases. Such pattern extraction facilitates evidence-based recommendations for undiagnosed or ambiguous cases, thereby enhancing decision reliability. This study introduces an adaptive fuzzy rule–based diagnostic framework for coronary artery disease (CAD) prediction, designed to process heterogeneous clinical datasets and provide interpretable decision support. The model incorporates data preprocessing, hybrid feature selection, data-driven membership function generation, and automatic construction of weighted and biased fuzzy rules within a Mamdani inference engine. The framework was evaluated on three standard datasets—Cleveland, Hungarian, and Switzerland—from the UCI Heart Disease Repository using Accuracy, Precision, Recall, and F1-score as evaluation metrics. Experimental results demonstrated accuracies of 89%, 80.9%, and 94.59% respectively, with corresponding F1-scores of 89.34%, 72.13%, and 97.22%. These results confirm that the proposed model outperforms conventional classifiers such as SVM, LR, and KNN in both balanced and imbalanced data conditions. The integration of bias-adjusted and data-driven rule weighting enhances minority-class detection and ensures clinical interpretability, establishing the framework as a reliable and scalable tool for computer-aided diagnosis in healthcare systems.