An Effective Method for Association Rule Mining based on Transactional Matrix

Abstract

of this paper is to present a new method based on transactional matrix for finding association rules more efficiently. Apriori algorithm is one of the classical algorithms for finding association rules, but it has limitations of number of times database scanned is too large and number of candidate itemsets generated is large.To reduce these two limitations a new method tries to find the association rules directly from a matrix which is generated from the transactional database . Keywords— Apriori algorithm, Association rule, Frequent itemsets, Transactional matrix generated in pruning operation, by applying transaction tag method. Dongme Sun, Sheohue Teng and et.al (5) presented a new technique based on forward and reverse scan of database. It produces the frequent itemsets more efficiently if applied with certain satisfying conditions. Sixue Bai, Xinxi Dai (6) presented a method called P-matrix algorithm to generate the frequent itemsets. It is found that the P-Matrix algorithm is more efficient and fast algorithm than Apriori algorithm to generate frequent itemsets. In this paper, a new method based on transactional matrix is presented to find the association rules from the large transactional database. In this approach a transactional matrix is generated directly from the database and then frequent itemsets and support of each frequent itemset is generated directly from the transactional matrix . It is found that the new proposed approach finds the frequent itemsets more efficiently. The performance of new method is compared with that of Apriori algorithm with the help of an example.