高级检索

    基于动态最小支持度的增量频繁序列挖掘

    Dynamic Minimum Support Incremental Frequent Trajectory Mining Algorithm

    • 摘要: 在轨迹数据集有新增数据且最小支持度变更情况下,为了实现频繁轨迹集能够快速更新以及解决轨迹数据库占用大量存储空间的问题,提出基于动态最小支持度的增量频繁序列挖掘算法。该算法能够充分利用频繁轨迹集信息,在有新增轨迹数据加入原始轨迹数据集且最小支持度变更时,通过频繁轨迹序列与频繁1序列相连接生成候选序列,利用非频繁轨迹后缀子序列置信度来估计非频繁轨迹支持度,实现动态更新频繁项集,并且在挖掘频繁轨迹后不再需要保存原始轨迹数据。通过轨迹数据集的挖掘实验,验证了本文算法支持度估计的精度和算法的有效性。

       

      Abstract: With the popularization of smart devices and the continuous development of wireless communication technology, a large amount of trajectory data is recorded in wireless networks, and frequent item mining for trajectory data has become a research hotspot. In order to update the frequent trajectory set when new data is added to the trajectory database and the minimum support is changed, we propose the dynamic minimum support incremental updating PrefixSpan algorithm (DMSIU-PrefixSpan). A method is proposed in the algorithm to estimate the support of a sequence, in which the confidence of its suffix subsequence is approximated to replace its own confidence when estimating the support of a sequence. With this method, the sequence support can be calculated without scanning the database. DMSIU-PrefixSpan algorithm can make full use of the information of frequent trajectory set. When incremental trajectory data is added to the original trajectory database and the minimum support changes, the candidate sequence is generated by connecting frequent trajectory sequence and frequent 1 sequence, and the candidate sequence support is estimated using the support estimation method to discover the frequent sequences. The algorithm no longer needs to save the original trajectory data and achieves fast update of the frequent trajectory set. The accuracy of the support degree estimation and the effectiveness of the algorithm of this paper are verified by the mining experiments of the trajectory database.

       

    /

    返回文章
    返回