摘要
如何高效地挖掘频繁项集是关联规则挖掘的主要问题。该文根据集合论和矩阵理论,提出一种基于矩阵的频繁项集挖掘算法。该算法只需扫描数据库一次,就能把所有事务转化为矩阵的行,把所有项和项集转化为矩阵的列,在对矩阵操作时能一次性产生所有频繁项集,且当支持度阈值改变时无需重新扫描数据库。实验结果表明,该算法的挖掘效率高于Apriori算法。
How to mine the frequent itemsets efficiently is a main problem in association rule mining. According to the theory of congregation and matrix, a frequent itemsets mining algorithm based on matrix is proposed. Through scanning database only once, all transactions are transformed to be rows of matrix and all items and itemsets are transformed to be columns of matrix. This algorithm can one-off product all frequent itemsets, and need not rescan the database when support threshold value changes. Experimental results show the mining efficiency of this algorithm is higher than Apriori algorithm.
出处
《计算机工程》
CAS
CSCD
北大核心
2009年第1期84-86,共3页
Computer Engineering
基金
教育部科学技术研究基金资助重点项目(205014)
河北省教育厅科研计划基金资助项目(2006143)
