Improving Accuracy and Interpretability of Clinical Decision Support Systems through Possibilistic Constrained Evolutionary Optimization

In this paper we propose a computable approach to represent medical rules by means of Fuzzy Clinical Decision Support Systems by preserving both accuracy and interpretability. Usually, prediction accuracy of these systems goes to overlook their linguistic interpretability and, in order to simultaneously optimize those conflicting properties, multi-objective evolutionary algorithms are adopted. Differently, our proposal relies on the alternative assumption that the interpretability-accuracy tradeoff problem can be approached as a single-objective constrained optimization problem. In this spirit, a Differential Evolution algorithm, with a selection operator suitably adapted to the aim, is used for membership function tuning by maximizing accuracy and fulfilling several constraints for linguistic distinguish ability degrees - a semantic property of fuzzy sets with notable relevance for interpretability of fuzzy models - evaluated through Possibility measure. The proposed approach has been tested on the Vertebral Column Data set, a recent medical database publicly available, with results that confirm the effectiveness of our method.