Mining frequent patterns with multiple minimum supports using basic Apriori
Tiantian Xu, Xiangjun Dong · 2013
Mining frequent patterns with multiple minimum supports is an important generalization of the association-rule-mining problem, which was proposed by Liu et al. Instead of setting a single minimum support threshold for all items in basic Apriori, they allow users to specify different minimum supports to different items, and an Apriori-based algorithm, named MSapriori, is developed to mine all frequent patterns with these multiple minimum supports. MSapriori is different in several aspects with the basic Apriori and is not easier to understand than the basic Apriori. So in this paper, we propose an algorithm, named MSB_apriori, which uses basic Apriori to solve this problem. Then we compare MSB_apriori and MSapriori in runtime and space in details and find an optimized approach to MSB_apriori. Accordingly, an optimized MSB_apriori, named MSB_apriori+, is proposed. Experimental results on real-life datasets show that the MSB_apriori+ is much more efficient than MSB_apriori and close to MSapriori. The advantages of MSB_apriori+ lie in that 1) it may be more suitable than MSapriori in some real applications; and 2) it is easier to understand and can be used as a substitute.