摘要
传统关联规则算法中事务扫描的重复性以及最小支持度设定的不可靠性会导致计算最大频繁集运行效率低并产生大量冗余的、无趣的规则问题,提出一种改进的Apriori关联规则挖掘算法。对候选项目进行约减,引入兴趣度测量因子对提取的关联规则进行优化。数据实验结果表明,该方法可提高传统关联规则挖掘算法的效率,避免传统关联算法中扫描的重复性,对Web访问用户行为分析具有一定的指导意义。
High frequency in transaction scanning and unreliability in minimum support settlement of traditional association rules algorithm lead to poor performance in calculating maximal frequency set and thus bringing about a lot of redundant boring rule problems. To address this issue, an improved Apriori algorithm for the mining association rules was proposed. The candidate sets were reduced, and interestingness measure factors were introduced to optimize the extracted association rules. The data ex- peri-mental results demonstrate that the proposed method can improve the efficiency of the traditional algorithm and decrease the repeatability in scanning. It has certain instructiveness for analyzing Web logs＇ behavior.
出处
《计算机工程与设计》
北大核心
2017年第4期852-856,975,共6页
Computer Engineering and Design
基金
国家自然科学基金项目(60905066)
重庆邮电大学自然科学基金项目(A2009-03)
电子商务与现代物流重庆市高校市级重点实验室重点基金项目(ECML201403)
