A Fuzzy System Approach for Diabetes Classification

Diabetes is a top priority in the public health system, being a chronic disease which has been steadily increasing over the past few decades. According to the World Health Organization, by the year 2045, 9.9% of the adult population will be diagnosed with diabetes. This paper proposes a new fuzzy inference for diabetes classification. It works as a system that generates case-specific diabetes prediction based on patient data. Starting from a real data acquisition, Decision Tree and Fuzzy Unordered Rule Induction Algorithm were chosen to generate rules. Starting from these rules, the proposed approach gives a more accurate understanding of the main factors that lead to diabetes. Furthermore, the architecture of our method is designed based on the American Diabetes Association criteria and can be used as an effective diabetes classification tool before reaching a doctor.