基于滑动窗口模型的数据流加权频繁模式挖掘算法

The Data Stream Weighted Frequent Pattern Mining Algorithm Based on the Sliding Window Model

加权频繁模式挖掘比传统的频繁模式挖掘更加的具有实际意义,针对数据流中的数据只能扫描有限次的性质,提出了基于滑动窗口模型的数据流加权频繁模式挖掘方法WFP-SW,该算法中数据存储采用的是矩阵数据结构,通过矩阵之间的相关操作来产生加权频繁模式。实验结果显示,该算法在产生加权频繁模式的时候不产生冗余模式,比传统的频繁模式挖掘算法有更好的效率。

The weighted frequent pattern mining algorithm has more practical implication than traditional frequent pattern mining ones.According to the nature of data stream which can only scans database several times,this paper proposes a weighted frequent pattern algorithm based on the sliding window model（WFP-SW）.The algorithm applies matrix data structure to store data,and then produces the weighted frequent pattern through some relative operation between matrixes. Experiment results show that the algorithm can avoid redundant patterns in the process of producing the weighted frequent pattern, and this algorithm is more efficient than the traditional frequent pattern mining algorithms.