最简有效关联规则及其挖掘算法

Minimal Availability Association Rules and Mining Algorithm

传统关联规则挖掘算法往往会产生过多规则而难以被决策者所采用。针对该问题,文章从应用的角度提出了最简有效关联规则,其特点是采用以后项为导向的挖掘方式,同时追求规则前后项之间的相关性,在此基础上给出了一种最简有效关联规则挖掘算法。利用该算法得到的最简有效关联规则集包括的规则数量大为减少且能得出与全部有效关联规则集相同的决策,避免了大量的冗余挖掘及无效挖掘,提高了挖掘效率和应用效果。

Conventional mining algorithms often produce too many rules for decision markets to digest. Instead, the concept of minimal availability association rules is introduced from the aspect of apphcation in this paper. Minimal availability rule set, which includes rules with item as consequent and the correlation of items as the antecedent, can be used to derive the same availability decisions as other association rules without availability information loss; while the number of minimal availability rules is much less than of all rules. A mining availability algorithm without redundant rules is proposed and the mining efficiency is improved.