Diabetes risk stratification method based on fuzzy logic and bio-inspired meta-heuristics

This paper presents a system for diabetes risk stratification that combines fuzzy logic with two bio-inspired algorithms. The developed system takes as input a set of patients described by numerical and categorical features and generates fuzzy rules to classify them into groups according to their risk of having diabetes. To take into consideration the uncertainty from the input dataset, our system combines fuzzy logic techniques with bio-inspired algorithms and hierarchical classification. The system has been evaluated on Pima Indians data from UCI Machine Learning Repository.