减少候选项集的数据流高效用项集挖掘算法

大数据环境下高效用项集挖掘算法中过多的候选项集极大地降低了算法的时空效率,提出了一种减少候选项集的数据流高效用项集挖掘算法。首先,通过数据流中当前窗口的一次扫描建立一个全局树,并降低全局树中头表入口与节点的冗余效用值；然后,基于全局树生成候选模式,基于增长算法降低局部树的候选项集效用；最终,从候选模式中选出高效用模式。基于真实数据流的实验结果表明,本算法的时空效率与内存占用比均优于其他数据流的高效用模式挖掘算法。

High utility itemsets mining algorithm of data stream with reducing candidate itemsets

In the big data stream scenario, high utility pattern mining algorithm generates a lot of candidate itemsets and it reduces the efficiency of time and space of algorithm, a high utility itemsets mining algorithm of data stream with reducing candidate itemsets is proposed to resolve that problem. Firstly, a global tree is constructed through a single scan of the current window in a data stream, redundancy utilities in both entries of a header table and nodes in the tree are reduced in this stage; secondly, candidate patterns are generated from the constructed tree, the redundancy utilities of local tree are reduced by growth algorithm; lastly, it identifies a set of high utility patterns from the candidate patterns. Realistic data streams based experimental results show that the proposed algorithm performs better in efficiency of time and space and memory usage index than the other high utility pattern mining algorithm of data streams.