Abstract
Expert systems are widely used as decision support tools in healthcare, particularly for heart disease diagnosis, which involves complex symptom patterns and high uncertainty. This study develops a heart disease diagnostic expert system based on Case-Based Reasoning (CBR) and examines the system from an epistemological perspective. Expert knowledge was acquired from a cardiology specialist and formalized into a structured case base consisting of symptoms, weighted parameters, and diagnostic outcomes. The inference mechanism follows the retrieve–reuse–revise–retain cycle, with case similarity computed using the Nearest Neighbor method. The results show that the proposed system is able to generate initial diagnostic recommendations by identifying the most similar previous cases in a transparent and systematic manner. In the experimental evaluation, the system achieved the highest similarity score of 67.5% in diagnosing cardiomyopathy cases, indicating that the similarity-based retrieval mechanism is capable of identifying relevant past cases that can support preliminary clinical interpretation. From an epistemological standpoint, the system represents empirical and experience-based medical knowledge that is computationally formalized, where diagnostic conclusions are probabilistic rather than definitive. The system’s epistemic limitations are mainly related to its dependence on the completeness and quality of the case base and its inability to fully capture clinical intuition and causal reasoning. Nevertheless, the CBR-based approach enhances transparency and accountability, as the reasoning process can be traced through matched symptoms, assigned weights, and similarity values. Overall, the expert system functions as a reliable decision support tool that integrates intelligent system techniques with applied epistemology to support rational and explainable medical decision-making.
Keywords— expert system, case-based reasoning, heart disease diagnosis, epistemology, decision support system.