摘要
本文提出了一种高效挖掘关联规则算法。该算法采用矩阵和向量表示事务数据库,基于FP_growth算法模式增长思想,引入索引跳跃技术,最大化虚拟地压缩了事务数据库,而且不产生侯选集,极大地加速了搜索的速度,从而有效地提高了产生关联规则的效率。
An efficient algorithm for mining association rules is proposed. The algorithm utilizes matrix and vector to indicate transaction databases. Based on the pattern-growth idea of the FP_growth algorithm, the algorithm introduces the index jumping technology, maximizing virtually the compression of transaction databases without generating candidate sets, which accordingly increases the speed of search and the efficiency of generating association rules.
出处
《计算机工程与科学》
CSCD
北大核心
2009年第4期69-71,共3页
Computer Engineering & Science
关键词
数据挖掘
索引
关联规则
条件模式
data mining
index
association rule
conditional pattern
