基于有向滑动时间窗的报警动态选择算法

Alert dynamic selection algorithm based on direct slide time window

针对现有报警选择方法在实时性和精确性方面存在的不足,提出一种基于有向滑动时间窗的报警动态选择算法。依据历史关联结果的趋向特征,动态调整每个时间窗内报警抽取的概率,向历史关联强度高的时间窗分配更高的抽取概率,提高报警关联的效率和准确性。仿真结果表明,该算法在实时性上远高于报警全选方法和随机选择方法,在关联精度上优于随机选择方法。

Aiming at the insufficiency of current alert selection approaches in terms of real-time and accuracy,a alert dynamic selection approach based on direct slide time window was proposed.The abstract probability of alerts from each time window were adj usted based on the tendency feature in the history,and higher abstract probability to the higher correlation strength in the history to improve the efficiency and accuracy of alert correlation.Experimental results show that the real time performance of the proposed algorithm is much higher than that of all selection approach and random selection approach,and its correlation accuracy is superior to that of random selection approach.