A Fuzzy Expert System for Early Diagnosis of Multiple Sclerosis

BACKGROUND: Artificial intelligence plays an important role in medicine. Specially, expert systems can be designed for diagnosis of disease. OBJECTIVE: Artificial intelligence can be used for diagnosis of disease. This study proposes an expert system for diagnosis of Multiple Sclerosis based on clinical symptoms and demographic characteristics. Specially, it recommends patients to refer to a specialist for further investigation. MATERIAL AND METHODS: In this empirical study, some symptoms of Multiple Sclerosis are mapped to fuzzy sets. Moreover, several rules are defined for prediction of Multiple Sclerosis. The fuzzy sets and rules form the knowledge base of the expert system. Patients enter their symptoms and demographic information via a user interface and Mamdani method is used in inference engine to produce the appropriate recommendation. RESULTS: The precision, recall, and F-measure are used as criteria to analyze the efficiency of the expert system. The results show that the designed expert system can recommend patients for further investigation as effective as specialists. Specially, while the proposed expert system recommended referring to a doctor for some healthy users, most of the MS patients are diagnosed. CONCLUSION: The proposed expert system in this study can analyze the symptoms of patients to predict the Multiple Sclerosis disease. Therefore, it can investigate initial status of patients in a rapid and cost-effective manner. Moreover, this system can be applied in situations and places, which human experts are unavailable.