基于加权Euclid范数的MTS异常检测

Outlier Detection of Multivariate Time Series Based on Weighted Euclid Norm

为了提高时间序列的异常检测算法的精度,根据主成分的累积贡献率选择序列及其主成分,在k_近邻局部离群点检测算法中采用加权Euclid范数距离作为k_近邻距离,从而实现对多变量时间序列的异常检测。为了验证算法的有效性,对测试数据进行了异常检测。实验结果表明,算法的精度和查全率比传统方法具有更大的优越性。

In order to improve the precision of anomaly detection algorithm for multivariate time series,based on the cumulative contribution rate of principal components,sequence and its principal components are selected,and in the k nearest neighbor local outlier detection algorithm,the weighted Euclid norm distance is used as the k nearest neighbor distance,so as to realize the anomaly detection of multivariate time series.In order to verify the effectiveness of the algorithm,anomaly detection was carried out on test data.The experimental results show that the algorithm has more advantages than traditional methods in the precision and recall.