The complexity and uncertainty of diagnostic information makes the diagnostic process difficult to learn, teach, and practice. Fuzzy logic methods, used successfully with complex industrial control problems, may be appropriate to model the range of uncertainties found in medical diagnostic information. A fuzzy systems model for use with diagnostic and other medical decisions is described.
... [Show full abstract] Combining a state space view of an animal in which each dimension of the space represents a variable of the animal with a fuzzy sets representation of the variables and states of the animal leads to a fuzzy systems model, which can be used to successfully diagnose disease. Partitioning the multidimensional state space of the animal into healthy and specific disease regions provides a diagnostic space for evaluating the health of the animal. When an input vector representing the variables of a sick animal is entered into the system, the model can provide a diagnosis and, potentially, a prognosis for that animal. The model can be implemented on a desktop computer for convenient use, and it provides a helpful geometric interpretation of the concepts of "diagnosis" and "prognosis" for teaching diagnostic reasoning. The fuzzy systems approach has advantages that are unavailable in other methods. The capability of fuzzy systems to act as universal approximators allows them to accommodate complex, nonlinear, imprecise, and even conflicting relationships to provide accurate knowledge representation. With these advantages over standard rule-based methods of modeling the medical diagnostic process, fuzzy expert systems have broad potential for use in medicine and warrant further study to determine their application and possible limitations.