Quickest anomaly detection: A case of active hypothesis testing

Kobi Cohen; Qing Zhao

doi:10.1109/ITA.2014.6804268

Quickest anomaly detection: A case of active hypothesis testing

Research output: Contribution to conference › Paper › peer-review

Abstract

The problem of quickest detection of an anomalous process among M processes is considered. At each time, a subset of the processes can be observed, and the observations follow two different distributions, depending on whether the process is normal or abnormal. The objective is a sequential search strategy that minimizes the expected detection time subject to an error probability constraint. This problem can be considered as a special case of active hypothesis testing first considered by Chernoff in 1959, where a randomized test was proposed and shown to be asymptotically optimal. For the special case considered in this paper, we show that a simple deterministic test achieves asymptotic optimality and offers better performance in the finite regime.

Original language	American English
DOIs	https://doi.org/10.1109/ITA.2014.6804268
State	Published - 1 Jan 2014
Externally published	Yes
Event	2014 IEEE Information Theory and Applications Workshop, ITA 2014 - San Diego, CA, United States Duration: 9 Feb 2014 → 14 Feb 2014

Conference

Conference	2014 IEEE Information Theory and Applications Workshop, ITA 2014
Country/Territory	United States
City	San Diego, CA
Period	9/02/14 → 14/02/14

Keywords

Sequential detection
dynamic search
hypothesis testing

All Science Journal Classification (ASJC) codes

Computer Science Applications
Information Systems

Access to Document

10.1109/ITA.2014.6804268

Cite this

@conference{79738350645a4413a2b86776b51fae0f,

title = "Quickest anomaly detection: A case of active hypothesis testing",

abstract = "The problem of quickest detection of an anomalous process among M processes is considered. At each time, a subset of the processes can be observed, and the observations follow two different distributions, depending on whether the process is normal or abnormal. The objective is a sequential search strategy that minimizes the expected detection time subject to an error probability constraint. This problem can be considered as a special case of active hypothesis testing first considered by Chernoff in 1959, where a randomized test was proposed and shown to be asymptotically optimal. For the special case considered in this paper, we show that a simple deterministic test achieves asymptotic optimality and offers better performance in the finite regime.",

keywords = "Sequential detection, dynamic search, hypothesis testing",

author = "Kobi Cohen and Qing Zhao",

year = "2014",

month = jan,

day = "1",

doi = "10.1109/ITA.2014.6804268",

language = "American English",

note = "2014 IEEE Information Theory and Applications Workshop, ITA 2014 ; Conference date: 09-02-2014 Through 14-02-2014",

}

TY - CONF

T1 - Quickest anomaly detection

T2 - 2014 IEEE Information Theory and Applications Workshop, ITA 2014

AU - Cohen, Kobi

AU - Zhao, Qing

PY - 2014/1/1

Y1 - 2014/1/1

N2 - The problem of quickest detection of an anomalous process among M processes is considered. At each time, a subset of the processes can be observed, and the observations follow two different distributions, depending on whether the process is normal or abnormal. The objective is a sequential search strategy that minimizes the expected detection time subject to an error probability constraint. This problem can be considered as a special case of active hypothesis testing first considered by Chernoff in 1959, where a randomized test was proposed and shown to be asymptotically optimal. For the special case considered in this paper, we show that a simple deterministic test achieves asymptotic optimality and offers better performance in the finite regime.

AB - The problem of quickest detection of an anomalous process among M processes is considered. At each time, a subset of the processes can be observed, and the observations follow two different distributions, depending on whether the process is normal or abnormal. The objective is a sequential search strategy that minimizes the expected detection time subject to an error probability constraint. This problem can be considered as a special case of active hypothesis testing first considered by Chernoff in 1959, where a randomized test was proposed and shown to be asymptotically optimal. For the special case considered in this paper, we show that a simple deterministic test achieves asymptotic optimality and offers better performance in the finite regime.

KW - Sequential detection

KW - dynamic search

KW - hypothesis testing

UR - http://www.scopus.com/inward/record.url?scp=84901236316&partnerID=8YFLogxK

U2 - 10.1109/ITA.2014.6804268

DO - 10.1109/ITA.2014.6804268

M3 - Paper

Y2 - 9 February 2014 through 14 February 2014

ER -

Quickest anomaly detection: A case of active hypothesis testing

Abstract

Conference

Keywords

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this