A novel pattern-frequency tree approach for transition analysis and anomaly detection in nonlinear and nonstationary systems

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The failure to identify anomaly patterns in dynamic systems can cause catastrophic events and incur a high cost. Prior research efforts attempted to use multiple sensors for a closer monitoring of the system dynamics. However, realizing full utilization of multiple sensors without the normality assumptions and dimensionality reduction remains a research challenge to build control schemes. This paper presents a novel methodology of pattern-frequency tree for transition analysis and anomaly detection in nonlinear and nonstationary systems. First, we propose Hyperoctree State space Aggregation Segmentation (HSAS) to delineate the high-dimensional dynamic processes in a continuous state space. Then, we develop a pattern-frequency tree to characterize and model the pattern distribution. Finally, we leverage pattern-frequency distribution information to develop a k-Maximin deviation algorithm for effective and efficient detection of process anomalies. Experimental results demonstrate that the proposed method performs better than the conventional methods in multi-sensor settings and high-dimensional environments.

Original languageEnglish (US)
Title of host publication67th Annual Conference and Expo of the Institute of Industrial Engineers 2017
PublisherInstitute of Industrial Engineers
Pages1264-1269
Number of pages6
ISBN (Electronic)9780983762461
StatePublished - 2017
Event67th Annual Conference and Expo of the Institute of Industrial Engineers 2017 - Pittsburgh, United States
Duration: May 20 2017May 23 2017

Other

Other67th Annual Conference and Expo of the Institute of Industrial Engineers 2017
CountryUnited States
CityPittsburgh
Period5/20/175/23/17

Fingerprint

Sensors
Dynamical systems
Agglomeration
Monitoring
Costs

All Science Journal Classification (ASJC) codes

  • Industrial and Manufacturing Engineering

Cite this

Chen, C. B., Yang, H., & Tirupatikumara, S. R. (2017). A novel pattern-frequency tree approach for transition analysis and anomaly detection in nonlinear and nonstationary systems. In 67th Annual Conference and Expo of the Institute of Industrial Engineers 2017 (pp. 1264-1269). Institute of Industrial Engineers.
Chen, Cheng Bang ; Yang, Hui ; Tirupatikumara, Soundar Rajan. / A novel pattern-frequency tree approach for transition analysis and anomaly detection in nonlinear and nonstationary systems. 67th Annual Conference and Expo of the Institute of Industrial Engineers 2017. Institute of Industrial Engineers, 2017. pp. 1264-1269
@inproceedings{36471f6c4421465192ab383c9c2b4ce7,
title = "A novel pattern-frequency tree approach for transition analysis and anomaly detection in nonlinear and nonstationary systems",
abstract = "The failure to identify anomaly patterns in dynamic systems can cause catastrophic events and incur a high cost. Prior research efforts attempted to use multiple sensors for a closer monitoring of the system dynamics. However, realizing full utilization of multiple sensors without the normality assumptions and dimensionality reduction remains a research challenge to build control schemes. This paper presents a novel methodology of pattern-frequency tree for transition analysis and anomaly detection in nonlinear and nonstationary systems. First, we propose Hyperoctree State space Aggregation Segmentation (HSAS) to delineate the high-dimensional dynamic processes in a continuous state space. Then, we develop a pattern-frequency tree to characterize and model the pattern distribution. Finally, we leverage pattern-frequency distribution information to develop a k-Maximin deviation algorithm for effective and efficient detection of process anomalies. Experimental results demonstrate that the proposed method performs better than the conventional methods in multi-sensor settings and high-dimensional environments.",
author = "Chen, {Cheng Bang} and Hui Yang and Tirupatikumara, {Soundar Rajan}",
year = "2017",
language = "English (US)",
pages = "1264--1269",
booktitle = "67th Annual Conference and Expo of the Institute of Industrial Engineers 2017",
publisher = "Institute of Industrial Engineers",
address = "United States",

}

Chen, CB, Yang, H & Tirupatikumara, SR 2017, A novel pattern-frequency tree approach for transition analysis and anomaly detection in nonlinear and nonstationary systems. in 67th Annual Conference and Expo of the Institute of Industrial Engineers 2017. Institute of Industrial Engineers, pp. 1264-1269, 67th Annual Conference and Expo of the Institute of Industrial Engineers 2017, Pittsburgh, United States, 5/20/17.

A novel pattern-frequency tree approach for transition analysis and anomaly detection in nonlinear and nonstationary systems. / Chen, Cheng Bang; Yang, Hui; Tirupatikumara, Soundar Rajan.

67th Annual Conference and Expo of the Institute of Industrial Engineers 2017. Institute of Industrial Engineers, 2017. p. 1264-1269.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - A novel pattern-frequency tree approach for transition analysis and anomaly detection in nonlinear and nonstationary systems

AU - Chen, Cheng Bang

AU - Yang, Hui

AU - Tirupatikumara, Soundar Rajan

PY - 2017

Y1 - 2017

N2 - The failure to identify anomaly patterns in dynamic systems can cause catastrophic events and incur a high cost. Prior research efforts attempted to use multiple sensors for a closer monitoring of the system dynamics. However, realizing full utilization of multiple sensors without the normality assumptions and dimensionality reduction remains a research challenge to build control schemes. This paper presents a novel methodology of pattern-frequency tree for transition analysis and anomaly detection in nonlinear and nonstationary systems. First, we propose Hyperoctree State space Aggregation Segmentation (HSAS) to delineate the high-dimensional dynamic processes in a continuous state space. Then, we develop a pattern-frequency tree to characterize and model the pattern distribution. Finally, we leverage pattern-frequency distribution information to develop a k-Maximin deviation algorithm for effective and efficient detection of process anomalies. Experimental results demonstrate that the proposed method performs better than the conventional methods in multi-sensor settings and high-dimensional environments.

AB - The failure to identify anomaly patterns in dynamic systems can cause catastrophic events and incur a high cost. Prior research efforts attempted to use multiple sensors for a closer monitoring of the system dynamics. However, realizing full utilization of multiple sensors without the normality assumptions and dimensionality reduction remains a research challenge to build control schemes. This paper presents a novel methodology of pattern-frequency tree for transition analysis and anomaly detection in nonlinear and nonstationary systems. First, we propose Hyperoctree State space Aggregation Segmentation (HSAS) to delineate the high-dimensional dynamic processes in a continuous state space. Then, we develop a pattern-frequency tree to characterize and model the pattern distribution. Finally, we leverage pattern-frequency distribution information to develop a k-Maximin deviation algorithm for effective and efficient detection of process anomalies. Experimental results demonstrate that the proposed method performs better than the conventional methods in multi-sensor settings and high-dimensional environments.

UR - http://www.scopus.com/inward/record.url?scp=85031003239&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85031003239&partnerID=8YFLogxK

M3 - Conference contribution

SP - 1264

EP - 1269

BT - 67th Annual Conference and Expo of the Institute of Industrial Engineers 2017

PB - Institute of Industrial Engineers

ER -

Chen CB, Yang H, Tirupatikumara SR. A novel pattern-frequency tree approach for transition analysis and anomaly detection in nonlinear and nonstationary systems. In 67th Annual Conference and Expo of the Institute of Industrial Engineers 2017. Institute of Industrial Engineers. 2017. p. 1264-1269