基于采样特异性因子的实时异常检测

Real-time Anomaly Detection Based on Sampled Peculiarity Factor

面向特异性的数据挖掘中,特异性因子是一个重要概念,但其计算时间复杂度过高。使用基于采样的方法定义特异性因子即采样特异性因子(Sampled Peculiarity Factor,SPF)可在不影响精度的情况下,提高运行效率。为提高基于SPF算法的异常检测效率,提出了基于SPF的学习采样频率算法,将SPF和最优采样频率结合起来提出了实时异常检测算法。在真实数据集上进行了实验,置信度为95%时,得到的最优采样频率序列为[1/32,1/16]。仿真实时异常实验表明该算法的误检率为2%。

Peculiarity factor is an important concept in the peculiarity-oriented mining, but its computation is too com- plex. Using sampling-based method to define the peculiarity factors called sampled peculiarity factor （SPF） can improve operational efficiency without affecting the accuracy. To improve the efficiency of anomaly detection algorithm based on the SPF, learning sampling frequency algorithm was proposed. Combined the optimal sampling frequency and SPF, real- time anomaly detection algorithm was proposed. Experiments use real data sets, take confidence as 95%, and the opti- mal sampling frequency sequence is between 1/32 and 1/16. Simulation results show that false detection rate of the al- gorithm is 2%.