摘要
在分析了频繁序列模式更新算法关键技术的基础上,提出了一种快速的增量式更新频繁序列模式挖掘算法FUFSPA,该算法将充分利用先前挖掘过程中所产生的信息来减少本次挖掘过程中的时间开销.另外,针对频繁序列模式挖掘中支持数计算的复杂性,提出了一种基于二进制形式的支持数计算方法,该方法只需进行一些“或”逻辑运算操作,将该方法用于序列模式挖掘中支持度(数)的计算,可以进一步提高算法的执行效率.实验结果表明算法FUFSPA是可行和有效的.
The key technique of updating frequent sequential patterns is studied. An incremental updating algorithm, FUFSPA, for mining frequent sequential patterns is presented, which makes use of information collection during an earlier mining process to cut down the cost of mining new sequential patterns in the updated database. Meanwhile, in order to solve the complexity of counting the support in mining frequent sequential patterns, an efficient method to calculate the support is proposed, which only executes some logical operation. The experiments show that the algorithm FUFSPA is efficient.
出处
《江苏大学学报(自然科学版)》
EI
CAS
北大核心
2007年第3期250-253,共4页
Journal of Jiangsu University:Natural Science Edition
基金
国家自然科学基金资助项目(60572112)
关键词
数据挖掘
频繁项目集
序列模式
增量式更新
关联规则
data mining
frequent itemsets
sequential pattern
incremental updating
association rules