A Belief Rule Based Decision Support System to Assess Multiple Disease Suspicion from Signs and Symptoms Under Uncertainty

This paper introduces a novel approach to decision support systems that have the capacity to diagnose multiple diseases, addressing the challenge of distinguishing between bronchiolitis and bronchopneumonia. These acute viral infections, prevalent in both children and older adults globally, present overlapping symptoms that can confound traditional diagnostic methods. To navigate this complexity, we employ the Belief Rule-Based Inference Methodology using Evidential Reasoning (RIMER), capable of handling diverse and maximal uncertainties in knowledge representation and inference. Our objective is to develop a decision support system, the Belief Rule-Based Expert System (BRBES), which integrates actual patient data and expert opinions into its knowledge base. Through rigorous testing on simulated patient data, our system demonstrates enhanced diagnostic performance, surpassing traditional methods as corroborated by Receiver Operating Characteristics Curve (ROC) analysis, thus establishing its reliability and superiority in multiple disease assessment.