Data Discovery and Anomaly Detection Using Atypicality for Real-Valued Data

Elyas Sabeti; Anders Høst-Madsen

doi:10.3390/e21030219

Data Discovery and Anomaly Detection Using Atypicality for Real-Valued Data

Entropy (Basel). 2019 Feb 26;21(3):219. doi: 10.3390/e21030219.

Authors

Elyas Sabeti¹, Anders Høst-Madsen^{2

3}

Affiliations

¹ Department of Computational Medicine and Bioinformatics, University of Michigan, NCRC 10-A108, 2800 Plymouth Rd, Ann Arbor, MI 48109-2800, USA.
² Department of Electrical Engineering, University of Hawaii at Manoa, Honolulu, HI 96822, USA.
³ Shenzhen Research Institute of Big Data, Shenzhen 518172, China.

Abstract

The aim of using 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 current paper extends this to real-valued data. This is done by using minimum description length (MDL). We develop the information-theoretic methodology for a number of "universal" signal processing models, and finally apply them to recorded hydrophone data and heart rate variability (HRV) signal.

Keywords: atypicality; big data; codelength; minimum description length.

Abstract

Grants and funding