基于人工蜂群的业务流异常状态检测方法

针对日益严重的网络安全问题,基于人工蜂群与聚类方法提出一种新的状态检测算法--DASA。该算法首先根据SKETCH方法和Hash函数建立业务流异常状态模型,并且利用人工蜂群技术实现对异常状态的检测。最后,以实际数据进行仿真实验,对比分析了样本数据与DASA算法检测的结果,发现DASA具有较好的适应性,而且聚类个数、丢弃阈值和邻域半径等因素对状态检测产生较大影响。

Detection method of anomaly traffic state based on artificial bee colony

In order to deal with the worsening network security problem, a new state detection algorithm, detection method of Anomaly traffic State based-Artificial bee colony (DASA), was proposed by Artificial Bee Colony (ABC) and clustering. In this algorithm, the anomaly traffic model was presented with SKETCH and Hash function at first, and the anomaly state was detected based on ABC. Then, a simulation with actual data is conducted to compare the results between Sample and DASA, which shows that DASA has better adaptability. And it has large impact on state detection with clustering number, dropping threshold and domain radius.