基于二进制的长频繁项目集挖掘算法

Algorithm of long frequent item sets mining based on binary

结合挖掘长频繁项目集的自顶向下搜索策略,提出一种基于二进制的长频繁项目集挖掘算法。该算法用数值递减搜索策略产生候选项,在用到频繁项目集修剪其子集减少候选项的基础上还通过事务特征减少搜索事务数,并运用二进制的逻辑"与"运算计算支持数,提高了算法的效率。算法分析和实验表明,该算法是有效的、快速的。

A long frequent item sets mining algorithm based on binary is presented, which combines top-down search that finds long frequent items sets. The algorithm uses descending-value search to create candidate-items, which are reduced by pruning subsets of fre- quent item sets, and further reduces searching the number of transaction by character of transactions, and then counts support with binary logic ＂and＂ operation, the efficiency of which is improved. Analysis and experiment of algorithm shows that it is efficient and fast.