Fuzzy confirmation measures

This paper proposes a number of measures for assessing the quality of fuzzy association rules. The focus is on a particular class of measures which are fuzzy counterparts of probabilistic confirmation measures. The distinguishing feature of such measures is that they are based on probabilistic (in)dependence and so a corresponding definition of fuzzy (in)dependence is proposed. This forms the basis for the definition of a fuzzy confirmation measure and for proposing a number of such measures based on a well-known fuzzy confidence measure. Various properties are considered to narrow the field of viable fuzzy confirmation measures, with one particular measure being found to have a number of advantages.