【摘要】 电信网络中各种频繁告警事件的发生通常具有一定的时序关系,据此提出了一种具有时序特征的告警关联挖掘算法。该方法解决了以往算法中无法发现告警序列之间时序关系的问题。其核心思想是在建立FP-tree时加入告警的时间特征,然后对FP-tree进行挖掘。实验证明,该算法能够快速、准确地从海量数据中挖掘出具有时序特征地关联规则。更多还原

【Abstract】 The frequent alarms in telecommunication networks usually have time-series correlation. To solve the deficient time-series correlation of nowadays algorithms, this paper introduces a time-series correlation data mining algorithm. It setups FP-tree by using time series to mine the rules. Experimental results show the algorithm can mine time-series association rules effectively.更多还原