Mining diagnostic rules using fuzzy clustering

To emphasize the importance of interpretability in fuzzy medical diagnosis, in this chapter we describe a framework, which enables the extraction of transparent diagnostic rules through fuzzy clustering. The methodology underlying the framework relies on two clustering steps. In the first step, a clustering algorithm is applied to the multi-dimensional data, in order to discover the hidden relationships among data. The second clustering step operates at the level of each input dimension to enable the definition of interpretable fuzzy sets to be used in the definition of diagnostic fuzzy rules. Differently from standard clustering schemes, this approach is able to extract fuzzy rules that satisfy interpretability constraints, so that a linguistic interpretation of diagnostic rules is immediate. Moreover, the framework is quite general and does not depend on specific clustering algorithms, which could be chosen according to different needs