arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2209.12398cs.AI
收藏

多变量数据流的实时的异常检测

Real-time Anomaly Detection for Multivariate Data Streams

Microsoft Corporation(微软公司)

Kenneth Odoh

摘要:

提出了一种基于概率指数加权移动平均(PEWMA)的实时的【形】多元数据流异常检测算法。我们的配方是弹性的(突然的瞬态,突然的分布,逐渐分布)的数据变化。新的异常检测例程利用增量在线算法来处理流。此外,我们提出的异常检测算法以无监督的方式工作,消除了对标记示例的需要。我们的算法表现良好,是弹性的概念漂移。

英文摘要:

We present a real-time multivariate anomaly detection algorithm for data streams based on the Probabilistic Exponentially Weighted Moving Average (PEWMA). Our formulation is resilient to (abrupt transient, abrupt distributional, and gradual distributional) shifts in the data. The novel anomaly detection routines utilize an incremental online algorithm to handle streams. Furthermore, our proposed anomaly detection algorithm works in an unsupervised manner eliminating the need for labeled examples. Our algorithm performs well and is resilient in the face of concept drifts.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑