基于广义后缀树的事件流频繁情节在线挖掘算法

现有的事件序列频繁情节挖掘多采用Apriori—like算法，此方法无法应用于事件流数据发掘。针对采用滑动窗口的事件流频繁情节发现问题，提出一种广义后缀树结构，在新事件加入时对情节树进行动态维护：为提高时空效率，采用频繁情节发生列表逐层构建的方法实现对搜索空间的划分．通过监控边界情节以发现频繁情节的变化。实验结果表明了算法的有效性和优越性。

Online Algorithm for Mining Frequent Episodes From Event Streams Based on Generalized Suffix Tree

Most state-of-art frequent episodes mining algorithm are apriori-like which not suitable for event streams mining. This paper considers the problem of mining frequent episodes on event streams over a sliding window. An novel generalized suffix-tree structure is proposed to discover and store frequent episodes. It maintains the data structure dynamically when appending new events. To improve spatio-temporal efficiency, the occurrence list of frequent episodes is used layer upon layer for dividing searching space and the bound episodes are monitored for finding changes of frequent episodes. The experimental result shows the efficiency and superiority of the proposed algorithm.