Abstract—The main aim of this paper is to present a new method based on transactional matrix and transaction reduction for finding frequent itemsets more efficiently. The association rule mining is based mainly on discovering frequent itemsets. Apriori algorithm is the most classical algorithm in association rule mining, but it has two fatal deficiencies: generation of a large number of candidate itemsets and scanning the database too many times. Apriori and other popular association rule mining algorithms mainly generate a large number of candidate 2-itemsets. To remove these deficiencies, a new method named Matrix Based Algorithm with Tags (MBAT) is proposed in this paper which finds the frequent itemsets directly from the transactional matrix which is generated from the database to generate association rules. Proposed algorithm greatly reduces the number of candidate itemsets, mainly candidate 2-itemsets
—Apriori algorithm, Association rule, Frequent itemsets, Transactional matrix, Transaction reduction
The authors are with the Department of Computer Science and Engineering, National Institute of Technology, Jalandhar, India (e-mail: firstname.lastname@example.org; email@example.com).
Cite: Harpreet Singh and Renu Dhir, "A New Efficient Matrix Based Frequent Itemset Mining Algorithm with Tags," International Journal of Future Computer and Communication
vol. 2, no. 4 pp. 355-358, 2013.