Local outlier detection for data streams in sensor networks

revisiting the utility problem invited paper

Mahsa Salehi, Christopher Leckie, James C Bezdek, Tharshan Vaithianathan

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

Abstract

Outlier detection is an important task in data mining, with applications ranging from intrusion detection to human gait analysis. With the growing need to analyze high speed data streams, the task of outlier detection becomes even more challenging as traditional outlier detection techniques can no longer assume that all the data can be stored for processing. While researchers mostly focus on detecting global outliers for data streams, detecting local outliers on streaming data has been neglected. This is an example of the utility problem in machine learning, where the machine learning algorithm needs to consider how the scarcity of a critical resource in the deployment environment affects the utility of any learned model. In this paper we focus on local outliers and propose an incremental solution assuming finite memory available. Our experimental results on a variety of data sets show that our solution is well suited to application environments with limited memory (e.g., wireless sensor networks) where the state of the system is changing.

Original languageEnglish
Title of host publication2015 IEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing [ISSNIP]
Subtitle of host publication7 -9 April 2015 Singapore
EditorsYu-Chee Tseng, Hongyi Wu, Lawrence Wai-Choong Wong
Place of PublicationPiscataway NJ USA
PublisherIEEE, Institute of Electrical and Electronics Engineers
Number of pages6
ISBN (Electronic)9781479980550, 9781479980543
DOIs
Publication statusPublished - 2015
Externally publishedYes
EventInternational Conference on Intelligent Sensors, Sensor Networks and Information Processing 2015 - , Singapore
Duration: 7 Apr 20159 Apr 2015
Conference number: 10th
https://web.archive.org/web/20151014113419/http://issnip2015.org/

Conference

ConferenceInternational Conference on Intelligent Sensors, Sensor Networks and Information Processing 2015
Abbreviated titleISSNIP 2015
CountrySingapore
Period7/04/159/04/15
Internet address

Cite this

Salehi, M., Leckie, C., Bezdek, J. C., & Vaithianathan, T. (2015). Local outlier detection for data streams in sensor networks: revisiting the utility problem invited paper. In Y-C. Tseng, H. Wu, & L. W-C. Wong (Eds.), 2015 IEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing [ISSNIP] : 7 -9 April 2015 Singapore [7106978] Piscataway NJ USA: IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ISSNIP.2015.7106978
Salehi, Mahsa ; Leckie, Christopher ; Bezdek, James C ; Vaithianathan, Tharshan. / Local outlier detection for data streams in sensor networks : revisiting the utility problem invited paper. 2015 IEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing [ISSNIP] : 7 -9 April 2015 Singapore. editor / Yu-Chee Tseng ; Hongyi Wu ; Lawrence Wai-Choong Wong. Piscataway NJ USA : IEEE, Institute of Electrical and Electronics Engineers, 2015.
@inproceedings{842854c923f84aed87b2eebee1c921f7,
title = "Local outlier detection for data streams in sensor networks: revisiting the utility problem invited paper",
abstract = "Outlier detection is an important task in data mining, with applications ranging from intrusion detection to human gait analysis. With the growing need to analyze high speed data streams, the task of outlier detection becomes even more challenging as traditional outlier detection techniques can no longer assume that all the data can be stored for processing. While researchers mostly focus on detecting global outliers for data streams, detecting local outliers on streaming data has been neglected. This is an example of the utility problem in machine learning, where the machine learning algorithm needs to consider how the scarcity of a critical resource in the deployment environment affects the utility of any learned model. In this paper we focus on local outliers and propose an incremental solution assuming finite memory available. Our experimental results on a variety of data sets show that our solution is well suited to application environments with limited memory (e.g., wireless sensor networks) where the state of the system is changing.",
author = "Mahsa Salehi and Christopher Leckie and Bezdek, {James C} and Tharshan Vaithianathan",
year = "2015",
doi = "10.1109/ISSNIP.2015.7106978",
language = "English",
editor = "Yu-Chee Tseng and Hongyi Wu and Wong, {Lawrence Wai-Choong}",
booktitle = "2015 IEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing [ISSNIP]",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
address = "United States of America",

}

Salehi, M, Leckie, C, Bezdek, JC & Vaithianathan, T 2015, Local outlier detection for data streams in sensor networks: revisiting the utility problem invited paper. in Y-C Tseng, H Wu & LW-C Wong (eds), 2015 IEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing [ISSNIP] : 7 -9 April 2015 Singapore., 7106978, IEEE, Institute of Electrical and Electronics Engineers, Piscataway NJ USA, International Conference on Intelligent Sensors, Sensor Networks and Information Processing 2015, Singapore, 7/04/15. https://doi.org/10.1109/ISSNIP.2015.7106978

Local outlier detection for data streams in sensor networks : revisiting the utility problem invited paper. / Salehi, Mahsa; Leckie, Christopher; Bezdek, James C; Vaithianathan, Tharshan.

2015 IEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing [ISSNIP] : 7 -9 April 2015 Singapore. ed. / Yu-Chee Tseng; Hongyi Wu; Lawrence Wai-Choong Wong. Piscataway NJ USA : IEEE, Institute of Electrical and Electronics Engineers, 2015. 7106978.

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

TY - GEN

T1 - Local outlier detection for data streams in sensor networks

T2 - revisiting the utility problem invited paper

AU - Salehi, Mahsa

AU - Leckie, Christopher

AU - Bezdek, James C

AU - Vaithianathan, Tharshan

PY - 2015

Y1 - 2015

N2 - Outlier detection is an important task in data mining, with applications ranging from intrusion detection to human gait analysis. With the growing need to analyze high speed data streams, the task of outlier detection becomes even more challenging as traditional outlier detection techniques can no longer assume that all the data can be stored for processing. While researchers mostly focus on detecting global outliers for data streams, detecting local outliers on streaming data has been neglected. This is an example of the utility problem in machine learning, where the machine learning algorithm needs to consider how the scarcity of a critical resource in the deployment environment affects the utility of any learned model. In this paper we focus on local outliers and propose an incremental solution assuming finite memory available. Our experimental results on a variety of data sets show that our solution is well suited to application environments with limited memory (e.g., wireless sensor networks) where the state of the system is changing.

AB - Outlier detection is an important task in data mining, with applications ranging from intrusion detection to human gait analysis. With the growing need to analyze high speed data streams, the task of outlier detection becomes even more challenging as traditional outlier detection techniques can no longer assume that all the data can be stored for processing. While researchers mostly focus on detecting global outliers for data streams, detecting local outliers on streaming data has been neglected. This is an example of the utility problem in machine learning, where the machine learning algorithm needs to consider how the scarcity of a critical resource in the deployment environment affects the utility of any learned model. In this paper we focus on local outliers and propose an incremental solution assuming finite memory available. Our experimental results on a variety of data sets show that our solution is well suited to application environments with limited memory (e.g., wireless sensor networks) where the state of the system is changing.

UR - http://www.scopus.com/inward/record.url?scp=84933573882&partnerID=8YFLogxK

U2 - 10.1109/ISSNIP.2015.7106978

DO - 10.1109/ISSNIP.2015.7106978

M3 - Conference Paper

BT - 2015 IEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing [ISSNIP]

A2 - Tseng, Yu-Chee

A2 - Wu, Hongyi

A2 - Wong, Lawrence Wai-Choong

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - Piscataway NJ USA

ER -

Salehi M, Leckie C, Bezdek JC, Vaithianathan T. Local outlier detection for data streams in sensor networks: revisiting the utility problem invited paper. In Tseng Y-C, Wu H, Wong LW-C, editors, 2015 IEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing [ISSNIP] : 7 -9 April 2015 Singapore. Piscataway NJ USA: IEEE, Institute of Electrical and Electronics Engineers. 2015. 7106978 https://doi.org/10.1109/ISSNIP.2015.7106978