Mining Double-Connective Association Rules from Multiple Tables of Relational Databases

Single-dimensional association rule and multidimensional association rule in conventional association rule mining are single-connective association rules, because they have only one connective: ¿. In this paper, based on mutually-inversistic logic, an algorithm for double-connective association rule mining is proposed, mining the association rule in the form of student(Sno)¿^-1 course(Cno)/¿^-1 study(Sno, Cno), read "for all students Sno, there exists course Cno such that Sno study Cno", where ¿^-1 and /¿^-1 are connectives, Sno is the primary key of the table transformed from the entity Student, Cno is the primary key of the table transformed from the entity Course, (Sno, Cno) is the primary key of the table transformed from the binary relationship Study. A three-table relational database naturally embraces double-connective association rules.