A Novel Feature Selection Scheme for Energy Efficient Wireless Sensor Networks

Moh'D Alwadi; Girija Chetty

doi:10.1007/978-3-642-33065-0_28

A Novel Feature Selection Scheme for Energy Efficient Wireless Sensor Networks

Moh'D Alwadi, Girija Chetty

Information Systems

Research output: A Conference proceeding or a Chapter in Book › Conference contribution › peer-review

4 Citations (Scopus)

Abstract

Energy efficiency is a key issue in wireless sensor networks where the energy sources and battery capacity are very limited. In this paper we propose a novel pattern recognition based formulation for minimizing the energy consumption in wireless sensor networks. The proposed scheme involves an algorithm to rank and select the sensors from the most significant to the least, and followed by a naïve Bayes classification. Assuming that each feature represents a sensor in the wireless sensor network, various data sets with multiple features are considered to show that feature ranking and selection could play a key role for the energy management. We have examined Isolet, forest fires and ionosphere datasets from the UCI repository to emulate the wireless sensor network scenario. From our simulation results, we show that it is possible to achieve two important objectives using the proposed scheme: (1) Increase the lifetime of the wireless sensor network, by using optimal number of sensors, and (2) Manage sensor failures with optimal number of sensors without compromising the accuracy

Original language	English
Title of host publication	12th International Conference on Algorithms and Architectures for Para, ICA3PP 2012
Editors	Yang Xiang, Ivan Stojmenovic, Bernady O Apduhan, Guojun Wang, Koji Nakano, Albert Zomaya
Place of Publication	Berlin Heidelberg
Publisher	Springer
Pages	264-273
Number of pages	10
Volume	7440
ISBN (Print)	9783642330653
DOIs	https://doi.org/10.1007/978-3-642-33065-0_28
Publication status	Published - 2012
Event	12th International Conference on Algorithms and Architectures for Parallel Processing ICA3PP 2012: ICA3PP 2012 - Fukuoka, Fukuoka, Japan Duration: 4 Sept 2012 → 7 Sept 2012 http://nsclab.org/ica3pp12/

Conference

Conference	12th International Conference on Algorithms and Architectures for Parallel Processing ICA3PP 2012
Abbreviated title	ICA3PP 2012
Country/Territory	Japan
City	Fukuoka
Period	4/09/12 → 7/09/12
Other	ICA3PP 2012 is the 12th in this series of conferences started in 1995 that are devoted to algorithms and architectures for parallel processing. ICA3PP is now recognized as the main regular event of the world that is covering the many dimensions of parallel algorithms and architectures, encompassing fundamental theoretical approaches, practical experimental projects, and commercial components and systems. As applications of computing systems have permeated in every aspects of daily life, the power of computing system has become increasingly critical. This conference provides a forum for countries around the world to exchange ideas for improving the computation power of computing systems. Following the traditions of the previous successful ICA3PP conferences held in Hangzhou, Brisbane, Singapore, Melbourne, Hong Kong, Beijing, Cyprus, Taipei, Busan, and Melbourne, ICA3PP 2012 will be held in Fukuoka, Japan. The objective of ICA3PP 2012 is to bring together researchers and practitioners from academia, industry and governments to advance the theories and technologies in parallel and distributed computing. ICA3PP 2012 will focus on two broad areas of parallel and distributed computing, i.e., architectures, algorithms and networks, and systems and applications. The conference of ICA3PP 2012 will be organized by Kyushu Sangyo University, Japan
Internet address	http://nsclab.org/ica3pp12/

Access to Document

10.1007/978-3-642-33065-0_28

Cite this

@inproceedings{29faea542ebc4fbf9be70ead6f4c810c,

title = "A Novel Feature Selection Scheme for Energy Efficient Wireless Sensor Networks",

abstract = "Energy efficiency is a key issue in wireless sensor networks where the energy sources and battery capacity are very limited. In this paper we propose a novel pattern recognition based formulation for minimizing the energy consumption in wireless sensor networks. The proposed scheme involves an algorithm to rank and select the sensors from the most significant to the least, and followed by a na{\"i}ve Bayes classification. Assuming that each feature represents a sensor in the wireless sensor network, various data sets with multiple features are considered to show that feature ranking and selection could play a key role for the energy management. We have examined Isolet, forest fires and ionosphere datasets from the UCI repository to emulate the wireless sensor network scenario. From our simulation results, we show that it is possible to achieve two important objectives using the proposed scheme: (1) Increase the lifetime of the wireless sensor network, by using optimal number of sensors, and (2) Manage sensor failures with optimal number of sensors without compromising the accuracy",

keywords = "wireless sensor networks, feature ranking, feature selection, data sets, accuracy, life time extension factor, WEKA machine learning framework",

author = "Moh'D Alwadi and Girija Chetty",

year = "2012",

doi = "10.1007/978-3-642-33065-0_28",

language = "English",

isbn = "9783642330653",

volume = "7440",

pages = "264--273",

editor = "Yang Xiang and Ivan Stojmenovic and Apduhan, {Bernady O} and Guojun Wang and Koji Nakano and Albert Zomaya",

booktitle = "12th International Conference on Algorithms and Architectures for Para, ICA3PP 2012",

publisher = "Springer",

address = "Netherlands",

note = "12th International Conference on Algorithms and Architectures for Parallel Processing ICA3PP 2012 : ICA3PP 2012, ICA3PP 2012 ; Conference date: 04-09-2012 Through 07-09-2012",

url = "http://nsclab.org/ica3pp12/",

}

Alwadi, MD & Chetty, G 2012, A Novel Feature Selection Scheme for Energy Efficient Wireless Sensor Networks. in Y Xiang, I Stojmenovic, BO Apduhan, G Wang, K Nakano & A Zomaya (eds), 12th International Conference on Algorithms and Architectures for Para, ICA3PP 2012. vol. 7440, Springer, Berlin Heidelberg, pp. 264-273, 12th International Conference on Algorithms and Architectures for Parallel Processing ICA3PP 2012, Fukuoka, Japan, 4/09/12. https://doi.org/10.1007/978-3-642-33065-0_28

A Novel Feature Selection Scheme for Energy Efficient Wireless Sensor Networks. / Alwadi, Moh'D; Chetty, Girija.
12th International Conference on Algorithms and Architectures for Para, ICA3PP 2012. ed. / Yang Xiang; Ivan Stojmenovic; Bernady O Apduhan; Guojun Wang; Koji Nakano; Albert Zomaya. Vol. 7440 Berlin Heidelberg: Springer, 2012. p. 264-273.

Research output: A Conference proceeding or a Chapter in Book › Conference contribution › peer-review

TY - GEN

T1 - A Novel Feature Selection Scheme for Energy Efficient Wireless Sensor Networks

AU - Alwadi, Moh'D

AU - Chetty, Girija

PY - 2012

Y1 - 2012

N2 - Energy efficiency is a key issue in wireless sensor networks where the energy sources and battery capacity are very limited. In this paper we propose a novel pattern recognition based formulation for minimizing the energy consumption in wireless sensor networks. The proposed scheme involves an algorithm to rank and select the sensors from the most significant to the least, and followed by a naïve Bayes classification. Assuming that each feature represents a sensor in the wireless sensor network, various data sets with multiple features are considered to show that feature ranking and selection could play a key role for the energy management. We have examined Isolet, forest fires and ionosphere datasets from the UCI repository to emulate the wireless sensor network scenario. From our simulation results, we show that it is possible to achieve two important objectives using the proposed scheme: (1) Increase the lifetime of the wireless sensor network, by using optimal number of sensors, and (2) Manage sensor failures with optimal number of sensors without compromising the accuracy

AB - Energy efficiency is a key issue in wireless sensor networks where the energy sources and battery capacity are very limited. In this paper we propose a novel pattern recognition based formulation for minimizing the energy consumption in wireless sensor networks. The proposed scheme involves an algorithm to rank and select the sensors from the most significant to the least, and followed by a naïve Bayes classification. Assuming that each feature represents a sensor in the wireless sensor network, various data sets with multiple features are considered to show that feature ranking and selection could play a key role for the energy management. We have examined Isolet, forest fires and ionosphere datasets from the UCI repository to emulate the wireless sensor network scenario. From our simulation results, we show that it is possible to achieve two important objectives using the proposed scheme: (1) Increase the lifetime of the wireless sensor network, by using optimal number of sensors, and (2) Manage sensor failures with optimal number of sensors without compromising the accuracy

KW - wireless sensor networks

KW - feature ranking

KW - feature selection

KW - data sets

KW - accuracy

KW - life time extension factor

KW - WEKA machine learning framework

U2 - 10.1007/978-3-642-33065-0_28

DO - 10.1007/978-3-642-33065-0_28

M3 - Conference contribution

SN - 9783642330653

VL - 7440

SP - 264

EP - 273

BT - 12th International Conference on Algorithms and Architectures for Para, ICA3PP 2012

A2 - Xiang, Yang

A2 - Stojmenovic, Ivan

A2 - Apduhan, Bernady O

A2 - Wang, Guojun

A2 - Nakano, Koji

A2 - Zomaya, Albert

PB - Springer

CY - Berlin Heidelberg

T2 - 12th International Conference on Algorithms and Architectures for Parallel Processing ICA3PP 2012

Y2 - 4 September 2012 through 7 September 2012

ER -

A Novel Feature Selection Scheme for Energy Efficient Wireless Sensor Networks

Abstract

Conference

Access to Document

Fingerprint

Cite this