Type 2 fuzzy representations of lung scans to predict pulmonary emboli

In this paper we describe the exploratory use of Type 2 fuzzy sets to represent the perceptions of lung scan images by experts in order to predict pulmonary emboli using Type 2 fuzzy relations. We used a GA to find suitable parameters for the fuzzy sets so that good classification was achieved. We present preliminary results with a limited data set demonstrating the potential power of the approach.