基于比特向量组的数据流邻近序列模式挖掘算法研究

Mining frequent continuous sequential patterns over data streams based on bit vector structure

引入项的半垂直比特向量结构,提出挖掘数据流邻近序列模式的MCSP-TSW算法.通过改进比特向量组结构和引入两个剪枝策略,提出改进的MCSP-TSW-Imp算法来减少判断一个候选序列是否频繁的时间.实验表明,两种算法空间消耗相当,但MCSP-TSW-Imp算法比MCSP-TSW算法具有较高的时间效率.

In this paper, we introduce the semi vertical bit vector structure, then propose a new al gorithm called MCSP TSW to mine frequent continuous sequential pattern from data stream. By im proving the structure and introducing two pruning strategies, an improved algorithm, called MCSP TSW Imp, is developed to further reduce the time of deciding whether a candidate sequence is fre quent or not. Experiments show that the memory usage of two algorithms is almost the same, but MCSP TSW Imp is outperforms than CSP SW in exaeutinn tim,=