Learning association rules and/or associative classification rules has been extensively studied in data mining and knowledge discovery community. Associative classification rules are considered as constrained association rules. Mining traditional association rules from transaction databases, however, has suffered some limitations. One of these limitations is that each transaction merely contains binary items with each item either present or absent in a transaction. Another limitation is that only positive association rules are discovered. Mining fuzzy association rules and discovering negative association rules have been developed to overcome these limitations, respectively. In this paper, we briefly introduce fuzzy association rules and negative association rules and especially discuss and compare what negative association rules look like. Approaches to discovering fuzzy association rules and negative association rules are combined to propose a new interestingness measure for both positive and negative fuzzy association rules. With this interestingness measure, an algorithm for mining those rules is described. Also, the special forms of association rules are presented, called associative classification rules. The algorithm to mining positive and negative fuzzy association rules is extended to learn positive and negative fuzzy associative classification rules.