关键词: 关联规则, 数据挖掘, 频繁模式树, 最大频繁项集, 超集检验

Abstract: The algorithms based on FP-tree, for mining maximal frequent patterns, have high performance but with many drawbacks. For example, they must recursively generate conditional FPtrees, have to do the process of superset checking. In order to overcome these drawbacks of the existing algorithms, an algorithm NonCheck Mining algorithm of Maximum Frequent Pattern (NCMFP）for mining maximal frequent patterns was put forward after the analysis of DMFIA-1 algorithm. In the algorithm, neither constructing conditional frequent pattern tree recursively nor superset checking was needed through modifying the structure of FP-tree. This algorithm reduced the number of mining through early prediction before mining. The application of a method to get the public intersection sets could obtain a complete result. The experiment shows that the efficiency of NCMFP is two to five times as much as that of the similar algorithms in the case of a relatively small support.

Key words: association rules, data mining, Frequent Pattern Tree（FP-tree）, maximal frequent itemsets, Superset Checking