基于Iceberg概念格的最大频繁项集挖掘

摘要

Some existing algorithms for mining Maximal Frequent Itemset(MFI) limit in scanning data sets frequently and tremendous candidate set size, etc. Based on Iceberg concept lattice model, this paper presents a maximal frequent itemsets mining algorithm Iceberg Concept Lattice Maxmal Frequent Itemset Algorithm(1CMFIA) in the Iceberg concept lattice. The algorithm builds the Iceberg concept lattice through .scanning the data sets at a time, by using the coverage relationship between frequent concepts in the Iceberg concept lattice, it can quickly calculate the maximum frequent concepts corresponding to the maximum frequent itemsets. The intension of all maximal frequent concepts is the set of all maximal frequent itemsets. Experimental results show that ICMFIA algorithm outperforms other existing algorithms in the number of scan data sets and mining efficiency.%最大频繁项集挖掘算法存在扫描数据集次数多和候选集规模过大等局限.基于Iceberg概念格模型,提出一种在Iceberg概念格上挖掘最大频繁项集的算法ICMFIA.该算法通过一次扫描数据集构建Iceberg概念格,利用Iceberg概念格中频繁概念之间良好的覆盖关系能快速计算出最大频繁项集所对应的最大频繁概念,所有最大频繁概念的内涵就是所求的最大频繁项集的集合.实验结果表明,该算法具有扫描数据集次数少和挖掘效率高的优点.