本文在分析以往各种算法的基础上，提出了一种适用于长模式挖掘的不产生候选集的

关联规则改进算法，我们利用数据库查询语言存储FP一tree的信息而不建立FP一tree结构，

提出约简属性和直接从频度表(f_list)中提取CLOSET的方法，从而节省时间和空间。

-Based on the analysis of the previous algorithms based on a model applicable to long-mining does not produce a candidate set of association rules to improve the algorithm, we use the database query language FP storage of information of a tree rather than the establishment of a tree structure of FP, proposed reduction properties and directly from the frequency table (f_list) method to extract CLOSET, thereby saving time and space.