摘要
本研究介绍了关联规则挖掘的基本概念,分析了经典的Apriori算法,提出一种改进的关联规则挖掘算法,解决了挖掘课程相关性关联规则的问题。改进算法的基本思想:①采用位图数据格式;②系统中会永久保留支持度为0的候选1项集和候选2项集,当系统需要运行时,首先采用数据库的过滤技术,可以很快得到频繁2项集。突破了这一瓶颈,系统运行速度将得到较大的提升。将该算法应用于课程相关性分析,实验结果表明改进的算法性能优于Apriori算法。
This paper introduces the basic concept of association rule, presents the traditional Apriori algorithm, and proposes an improved algorithm of mining association rules. The main idea of this improved algorithm is （1）it uses bitmap; （2）the candidate 1-itemsets and candidate 2- itemsets whose support are 0 can be kept in the system for ever. When we do mining, we first adopt the technique of Database Filter, which can get the Frequent 2-itemsets quickly. With the breaking of this bottleneck, the operation speed of the system can be increased significantly. This algorithm is applied to course relativity analysis. The experimental results show that this algorithm is better than Apriori algorithm.
出处
《河北农业大学学报》
CAS
CSCD
北大核心
2010年第3期116-119,共4页
Journal of Hebei Agricultural University
基金
国家自然科学基金资助项目(60603027)
河北省科技研究与发展指导计划项目(07213543)
天津市科技发展计划资助项目(04310941R)
天津市应用基础研究计划资助项目(05YFJMJC11700)
