Deriving concise association rules efficiently while extracting closed itemsets in depth-first manner

作者：Wang Bingzheng^*; Wu Huixin; Ma Jiming

来源：Journal of Convergence Information Technology, 2012, 7(10): 68-76.

DOI：10.4156/jcit.vol7.issue10.8

• Closed itemsets
• Concise representations
• Depth first
• Divide and conquer
• Frequent closed itemsets
• Item sets
• Itemset
• Mining associations
• Search spaces

Mining association rules consists of two phases. One is to find proper itemsets, the other is to generate specific rules. Integrating the two steps properly can lead to efficient results. In this paper, a new method of extracting frequent closed itemsets in depth-first manner is proposed. It can mine the closed itemsets directly. At the same time, specific concise association rules are filtered base on those closed itemsets generated. As the process goes on, we can derive all association rules in concise representation correctly when the algorithm is completed. Through exploiting the depth-first mining strategy, the two steps share same divide-and-conquer process and data structure, thus prune more search space and achieve good performance. Experimental results show that the algorithm runs properly and performs better than those traditional ones.

• 出版日期2012
• 单位郑州大学