Sensitive, specific, and interpretable: evolving a fuzzy mammographic-interpretation assessment tool

In this paper we focus on the problem of mammographic interpretation and in the construction of a computerized system to assess physicians in this task. We propose a solution based on fuzzy systems, and apply our fuzzy modeling methodology-Fuzzy CoCo-to find systems exhibiting both good performance and high interpretability. The results presented show the capacity of Fuzzy CoCo to find systems that satisfy different requirements concerning the complexity of the solution, expressed mainly as the number of rules