Mining Binary Diagram Rules

Mining association rules was introduced as a novel framework of Data Mining by Agrawal et al. [1] and some e cient heuristic algorithms for mining all association rules have been devised [2]. Based on such research, some systems for mining association rules have been already in commerce and have succeeded in discovering a rule like \if a customer buys a swimsuit and a pair of ippers then the customer buys goggles too in 90% of the cases" from huge databases. This method is obviously applicable to genomic data and a successful result has been reported for nding simple association rules from signals found in mammalian promoter sequences [4]. In order to increase the diversity and expressibility of rules, we introduce a rule in a more general form, Binary Diagram Rule (BDR, shortly), so that a rule like \at teatime, if a person