用自动模糊划分与改进Apriori算法生成QAR关联规则

针对属性粒度模糊划分需事先给定与Aprirori算法效率低的问题,提出基于自动模糊划分和改进Apriori算法的QAR关联规则生成方法。首先对QAR数据进行空缺值填补等预处理;然后给出最佳聚类准则并根据给出的最佳聚类准则得到最佳聚类,从而对QAR属性完成自动模糊划分及隶属函数的确定;之后通过记录数据项位置及简化连接与剪枝过程来提高Apriori算法的效率;并将其应用到QAR关联规则的生成过程;最后通过品质和性能度量两方面的实验,表明此方法在各方面的性能均优于经典方法。

Generation of QAR Association Using Rules Based on Automatic Fuzzy Partition and Improved Apriori Algorithm

Considering the problems that attribute granularity fuzzy partition should be given in advance and the Apriori algorithm has low efficiencywe proposed a method for generating association rules in Quick Access Recorder(QAR) based on automatic fuzzy partition and improved Apriori algorithm.Firstpreprocessing was made for QAR data by filling the vacancies value.Thenthe optimum clustering criteria was given and an optimum clustering was obtained according to the criteria.Thus the fuzzy partition of QAR attributes was implemented and membership function was determined.It improved the efficiency of the algorithm by recording the location of the data items and simplifying the process of the pruning and connection.It was applied in the generation of the QAR association rules.Experiment was made on quality and performance of the methodand the result showed that the method is superior to the classical method in all aspects of performance.