We present an overview of association criteria that build upon the relative position of a set of reference data items with respect to given query data items, and propose fuzzy generalizations that allows to use these criteria as real-valued similarity measures. Some experimental consistency tests are also presented.