摘要
关联规则挖掘中,往往需要生成频繁集,一般需要大量的系统资源,为降低这一开销,将位运算算法应用到频繁集的生成逻辑运算中。将连接、剪枝等转换成整形数组按位"与"运算进行,即高速的数字运算代替了低速的比较查找运算。算法从最大项集开始查找,与Apriori及一般传统的查找算法方向相反,在事务项集较多,频繁项集可能是大项集时,可以缩短查找时间。
Finding frequent itemsets is often used to find the association rules,which need great amount of system resource. In order to reduce the research costs,the bit operation was put into the frequent itemsets. The join and prune were transformed to 'and' operation,meaning that the high speed digital operation has taken place of the low speed compare-search operation. Another characteristic of this algorithm is searching from the biggest itemsets,just reversing the direction of Apriori and some traditional algorithms. It can decrease the search time when the tasks include many items and the frequent itemsets maybe big intemsets.
出处
《计算机应用》
CSCD
北大核心
2004年第11期88-90,共3页
journal of Computer Applications
关键词
频繁集
逻辑运算
与运算
最大项集
frequent itemsets
logic operation
and operation
the most itemsets