基于索引数组与集合枚举树的最大频繁项集挖掘算法

A Maximal Frequent Itemset Mining Algorithm Based on Index Array and Set-Enumeration Tree

由于其内在的计算复杂性,挖掘密集型数据集的全部频繁项集非常困难,解决方案之一是挖掘最大频繁项集。集合枚举树是最大频繁项集挖掘算法中常用的数据结构,最大频繁项集的挖掘过程也可以看作是集合枚举树的搜索过程。为缩小集合枚举树的搜索空间,采用宽度优先和深度优先相结合的混合搜索策略,提出了一种新的最大频繁项集的挖掘算法Index-MaxMiner。该算法首先设计了索引数组这种新的数据结构,并给出了一个基于二进制位图技术的索引数组的计算方法。通过为每个频繁项增加包含索引,Index-MaxMiner利用一次宽度优先搜索得到了候选最大频繁项集,使集合枚举树的第一层结点个数大幅度减少。然后在候选最大频繁项集中通过深度优先搜索,得到全部最大频繁项集,从而实现了集合枚举树的跳跃式搜索,大大缩小了搜索空间。实验结果表明,该算法可有效提高最大频繁项集的挖掘效率。

Because of the inherent computational complexity, mining the complete frequent itemset in dense datasets remains to be a challenging task. Mining Maximal Frequent Itemset （MFI）is an alternative to address the problem. Set-Enumeration Tree （SET）is a common data structure used in several MFI mining algorithms. For this kind of algorithrns, the process of mining MFI can also be viewed as the process of searching in SET. To reduce the search space of SET, in this paper, a new algorithm, Index-MaxMiner, for mining MFI is proposed by employing a hybrid search strategy blending breadth-first and depth-first combined. Firstly, the ＂index array＂ is proposed, and by using bitmap, an algorithm for computing index array is presented. By adding subsume index to frequent items, Index-MaxM/ner discovers the candidate MFIs using breadth-first search at one time. By doing so, the number of nodes in the first level of SET is reduced greatly. Then, for candidate MFIs, depth-first search strategy is used to generate all MFIs. Thus, the jumping search in SET is implemented, and the search space is reduced greatly. The experimental results show that the proposed algorithm outperforms similar state-of-the-art algorithms.