Novelty detection for time series data analysis in water distribution systems using support vector machines

Stephen R Mounce, Richard B Mounce, Joby B Boxall

Research output: Contribution to journalArticle

52 Citations (Scopus)

Abstract

The sampling frequency and quantity of time series data collected from water distribution systems has been increasing in recent years, giving rise to the potential for improving system knowledge if suitable automated techniques can be applied, in particular, machine learning. Novelty (or anomaly) detection refers to the automatic identification of novel or abnormal patterns embedded in large amounts of "normal" data. When dealing with time series data (transformed into vectors), this means abnormal events embedded amongst many normal time series points. The support vector machine is a data-driven statistical technique that has been developed as a tool for classification and regression. The key features include statistical robustness with respect to non-Gaussian errors and outliers, the selection of the decision boundary in a principled way, and the introduction of nonlinearity in the feature space without explicitly requiring a nonlinear algorithm by means of kernel functions. In this research, support vector regression is used as a learning method for anomaly detection from water flow and pressure time series data. No use is made of past event histories collected through other information sources. The support vector regression methodology, whose robustness derives from the training error function, is applied to a case study.

Original languageEnglish
Pages (from-to)672-686
Number of pages15
JournalJournal of Hydroinformatics
Volume13
Issue number4
DOIs
Publication statusPublished - Nov 2011

Fingerprint

Water distribution systems
Support vector machines
Time series
time series
anomaly
Learning systems
outlier
nonlinearity
Sampling
water flow
learning
data analysis
detection
water distribution system
support vector machine
methodology
sampling
Water

Keywords

  • data analysis
  • leakage
  • novelty detection
  • support vector machines
  • water distribution systems

ASJC Scopus subject areas

  • Atmospheric Science
  • Geotechnical Engineering and Engineering Geology

Cite this

Novelty detection for time series data analysis in water distribution systems using support vector machines. / Mounce, Stephen R; Mounce, Richard B; Boxall, Joby B.

In: Journal of Hydroinformatics, Vol. 13, No. 4, 11.2011, p. 672-686.

Research output: Contribution to journalArticle

Mounce, Stephen R ; Mounce, Richard B ; Boxall, Joby B. / Novelty detection for time series data analysis in water distribution systems using support vector machines. In: Journal of Hydroinformatics. 2011 ; Vol. 13, No. 4. pp. 672-686.
@article{b8f729f5918c4a6ea360df836a38d04b,
title = "Novelty detection for time series data analysis in water distribution systems using support vector machines",
abstract = "The sampling frequency and quantity of time series data collected from water distribution systems has been increasing in recent years, giving rise to the potential for improving system knowledge if suitable automated techniques can be applied, in particular, machine learning. Novelty (or anomaly) detection refers to the automatic identification of novel or abnormal patterns embedded in large amounts of {"}normal{"} data. When dealing with time series data (transformed into vectors), this means abnormal events embedded amongst many normal time series points. The support vector machine is a data-driven statistical technique that has been developed as a tool for classification and regression. The key features include statistical robustness with respect to non-Gaussian errors and outliers, the selection of the decision boundary in a principled way, and the introduction of nonlinearity in the feature space without explicitly requiring a nonlinear algorithm by means of kernel functions. In this research, support vector regression is used as a learning method for anomaly detection from water flow and pressure time series data. No use is made of past event histories collected through other information sources. The support vector regression methodology, whose robustness derives from the training error function, is applied to a case study.",
keywords = "data analysis, leakage, novelty detection, support vector machines, water distribution systems",
author = "Mounce, {Stephen R} and Mounce, {Richard B} and Boxall, {Joby B}",
year = "2011",
month = "11",
doi = "10.2166/hydro.2010.144",
language = "English",
volume = "13",
pages = "672--686",
journal = "Journal of Hydroinformatics",
issn = "1464-7141",
publisher = "IWA Publishing",
number = "4",

}

TY - JOUR

T1 - Novelty detection for time series data analysis in water distribution systems using support vector machines

AU - Mounce, Stephen R

AU - Mounce, Richard B

AU - Boxall, Joby B

PY - 2011/11

Y1 - 2011/11

N2 - The sampling frequency and quantity of time series data collected from water distribution systems has been increasing in recent years, giving rise to the potential for improving system knowledge if suitable automated techniques can be applied, in particular, machine learning. Novelty (or anomaly) detection refers to the automatic identification of novel or abnormal patterns embedded in large amounts of "normal" data. When dealing with time series data (transformed into vectors), this means abnormal events embedded amongst many normal time series points. The support vector machine is a data-driven statistical technique that has been developed as a tool for classification and regression. The key features include statistical robustness with respect to non-Gaussian errors and outliers, the selection of the decision boundary in a principled way, and the introduction of nonlinearity in the feature space without explicitly requiring a nonlinear algorithm by means of kernel functions. In this research, support vector regression is used as a learning method for anomaly detection from water flow and pressure time series data. No use is made of past event histories collected through other information sources. The support vector regression methodology, whose robustness derives from the training error function, is applied to a case study.

AB - The sampling frequency and quantity of time series data collected from water distribution systems has been increasing in recent years, giving rise to the potential for improving system knowledge if suitable automated techniques can be applied, in particular, machine learning. Novelty (or anomaly) detection refers to the automatic identification of novel or abnormal patterns embedded in large amounts of "normal" data. When dealing with time series data (transformed into vectors), this means abnormal events embedded amongst many normal time series points. The support vector machine is a data-driven statistical technique that has been developed as a tool for classification and regression. The key features include statistical robustness with respect to non-Gaussian errors and outliers, the selection of the decision boundary in a principled way, and the introduction of nonlinearity in the feature space without explicitly requiring a nonlinear algorithm by means of kernel functions. In this research, support vector regression is used as a learning method for anomaly detection from water flow and pressure time series data. No use is made of past event histories collected through other information sources. The support vector regression methodology, whose robustness derives from the training error function, is applied to a case study.

KW - data analysis

KW - leakage

KW - novelty detection

KW - support vector machines

KW - water distribution systems

UR - http://www.scopus.com/inward/record.url?scp=80054993033&partnerID=8YFLogxK

U2 - 10.2166/hydro.2010.144

DO - 10.2166/hydro.2010.144

M3 - Article

VL - 13

SP - 672

EP - 686

JO - Journal of Hydroinformatics

JF - Journal of Hydroinformatics

SN - 1464-7141

IS - 4

ER -