Fuzzy logic in medical expert systems

Although expert systems technology has progressed rapidly in recent years, many systems still lack the ability to reason with imprecise or uncertain knowledge. It is this aspect, which is a task easily accomplished in the human reasoning process, which prevents these systems from assuming a more prominent role in practical applications. In this paper, sources of uncertainty and imprecision are analyzed, as they occur in the knowledge base, in the reasoning process, and in the actual data for each individual case. Techniques for handling these uncertainties are presented as they are applied in the EMERGE medical expert system for analysis of chest pain.