arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1709.03191eess.SPcs.ITmath.IT
收藏

基于非典型性的数据发现与异常检测:信号处理方法

Data Discovery and Anomaly Detection Using Atypicality: Signal Processing Methods

Elyas Sabeti, Anders Høst-Madsen

摘要:

重复性的目的是从大数据中提取小的,罕见的,不寻常的和有趣的片段。这补充了有关典型数据的统计数据,以深入了解数据。为了找到这些“有趣”的数据部分,需要通用的方法,因为我们事先不知道我们在寻找什么。因此,我们基于代码长度的精确性标准。在以前的论文中,我们开发了离散值数据的方法,目前的论文将其扩展到实值数据。这是通过使用最小描述长度(MDL)来完成的。我们表明,这与离散值的情况下,共享一些理论性质。我们开发了一些“通用”的信号处理模型的方法,并最终将其应用于记录的水听器数据。

英文摘要:

The aim of atypicality is to extract small, rare, unusual and interesting pieces out of big data. This complements statistics about typical data to give insight into data. In order to find such "interesting" parts of data, universal approaches are required, since it is not known in advance what we are looking for. We therefore base the atypicality criterion on codelength. In a prior paper we developed the methodology for discrete-valued data, and the the current paper extends this to real-valued data. This is done by using minimum description length (MDL). We show that this shares a number of theoretical properties with the discrete-valued case. We develop the methodology for a number of "universal" signal processing models, and finally apply them to recorded hydrophone data.

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。

↑