摘要
针对Apriori算法进行多值属性关联规则挖掘时效率低下的问题,提出量化Apriori算法。利用多值属性数据特点改变项集存储格式,采用类似矩阵的数据结构存储项集,提高遍历数据库时统计计数的速度,使用类似矩阵的加法运算改进连接操作,减少无效候选项集的产生。实验结果表明,相比Apriori算法,该算法执行效率有较大提高。
Aiming at the problem that the Apriori algorithm is inefficient in quantitative association rules mining,this paper proposes a Quantitative Apriori(Q-Apriori) algorithm.It makes use of quantitative attribute data's characteristics,changes the storage format of the sets using special data structure like matrix to store the sets which reduces the time of traversing the database to count each set's support.It makes improvement in join step using a method like matrix addition which reduces the number of nonsense candidate sets.Experimental results show that execution efficiency of this algorithm is better than that of Apriori algorithm.
出处
《计算机工程》
CAS
CSCD
北大核心
2011年第9期81-83,共3页
Computer Engineering
基金
国家自然科学基金资助项目(50674086)
江苏省博士后科学基金资助项目(0701045B)
中国矿业大学科技基金资助项目(2007B017)
关键词
关联规则
多值属性
数据挖掘
量化Apriori算法
association rules
multivalue attribute
data mining
Quantitative Apriori(Q-Apriori) algorithm