Call for Papers

First International Workshop on Data Mining and Wireless Sensor Networks (DM-WSN)
December 18, Hong Kong

DM-WSN'06 is the First International Workshop Data Mining and Wireless Sensor Networks (DM-WSN), which is organized in conjunction with the IEEE International Conference on Data Mining, ICDM'06, Hong Kong, December 18 - 22, 2006.

Description

Due to recent advancement in electronics industry, Wireless Sensor Networks (WSNs) are used for various applications such as security, agriculture and environmental monitoring. WSN may contain hundreds of tiny, low-cost, battery-powered devices that monitor physical attributes (humidity, temperature, and light) and that self-organize into networks that can make autonomous decisions (turn on/off actuators), and are part of a larger distributed management and control system (e.g. irrigation system). As each node is a data source, a sensor network can possibly generate large sets of data, which are ideal candidates for data mining techniques. However, sensor networks are constrained in their ability to communicate their data to a centralized processing server where data mining would normally take place. Sensors are limited in terms of available energy for transmission, computational power, memory, and communications bandwidth. Distributed data mining (DDM) methods provide solutions to these constraints by placing aspects of the data mining process such as data sampling, aggregation, and modeling on individual sensors, as well as clusters of sensors. These activities and placement in the sensor network vary by the type of data mining being undertaken such as classification, prediction, time series analysis, clustering, and anomaly detection. The focus of this workshop is to share the lessons learned from previous successful applications of DDM for WSNs and to discuss new theories to distribute the data mining process over large sensor networks. We are particularly interested in approaches that have solved global network data mining problems through localized and distributed computation.

Contribution

This workshop will bring together researchers and practitioners from academia and industry. The workshop objectives are as follows:
  • Collect and disseminate lessons learned from prior applications of DDM to sensor networks - business, science, engineering, medicine, and other disciplines.
  • Present and discuss new theoretical results, innovative ideas, and preliminary studies on DDM that allow knowledge discovery to scale to sensor networks of massive size.
  • Overcome the individual sensor constraints of available energy for transmission, computational power, memory, and communications bandwidth, so as to more efficiently undertake the data mining process on the sensor network.

Potential Topics:

  • Power consumption characteristics of distributed data mining algorithms and developing data mining algorithms that minimize power consumption.
  • DDM methods that overcome sensor limitations such as available energy for transmission, computational power, memory, and communications bandwidth.
  • Efficient, scalable and distributed algorithms for large-scale DDM tasks such as classification, prediction, link analysis, time series analysis, clustering, and anomaly detection.
  • DDM methods that distribute aspects of the data mining process such as data selection, sampling, cleaning, reduction, transformation, integration and aggregation, as well as model development, validation and deployment.
  • Theory and application of:
  • Distributed Principal Component Analysis (PCA) and Independent Component Analysis (ICA)
  • Distributed machine learning (neural networks, support vector machines, decisions trees and rules, genetic algorithms)
  • Distributed statistical regression methods
  • Distributed Bayesian learning (belief networks, decision networks)
  • Distributed clustering methods (distributed k-Means, dynamic neural networks)
  • Agent based approaches to DDM.
  • Incremental, exploratory and interactive mining
  • Visual data mining
  • Theoretical foundations in DM and WSNs; extensions of computational learning theory to sensor networks
  • Mining of data streams
  • Comparisons of in-network DDM versus traditional server side DM
  • Privacy sensitive data mining
  • Successful applications of DM for WSN in business, science, engineering, medicine, and other disciplines with particular attention to lessons learned.

Workshop Organizers

Sajid Hussain
Jodrey School of Computer Science,
Acadia University,
Wolfville, NS, Canada
http://cs.acadiau.ca/~shussain/ Sajid.Hussain@acadiau.ca
Danny Silver
Jodrey School of Computer Science,
Acadia University,
Wolfville, NS, Canada
http://plato.acadiau.ca/courses/
comp/dsilver/

Danny.Silver@acadiau.ca
Qiang Yang
Department of Computer Science
Hong Kong University of Science and Technology
Kowloon, Hong Kong
http://www.cs.ust.hk/~qyang/
qyang@cs.ustk.hk
Workshop Website- http://cs.acadiau.ca/~shussain/dm-wsn

Technical Program Committee (TPC) - (confirmed)

NameAffiliation
Abidi, S.S. RazaDalhousie University, Canada
Bontempi, GianlucaUniversite Libre de Bruxelles, Belgium
Chen, LeiHong Kong Uni. of Science & Tech., Hong Kong
Chou, Cheng-fuNational Taiwan University, Taiwan
Graham, PeterUniversity of Manitoba, Canada
Hammad, Moustafa A.University of Calgary, Canada
Jamil, Hasan M.Wayne State University
Kemke, ChristelUniversity of Manitoba, Canada
Leung, Carson K.University of Manitoba, Canada
Pan, Jeffrey J.Hong Kong Uni. of Sci. & Tech., Hong Kong
Peng, Wen-ChihNational Chiao Tung University, Taiwan
Sander, JoergUniversity of Alberta, Canada
Talbi, El-GhazaliLIFL, Université de Lille, France
Tan, Pang-NingMichigan State University
Yang, Laurence T.St. Francis Xavier University, Canada
Yin, JieHong Kong Uni. of Science & Tech., Hong Kong
Zaki, MohammedRensselaer Polytechnic Institute

Important Dates

Submission deadlineJuly 30, 2006
Authors NotificationSeptember 08, 2006
Final ManuscriptSeptember 29, 2006
Workshop DayDecember 18, 2006

Paper Submission

IEEE Computer Society format, (4+1 extra) pages, check the Submission page on ICDM'06 website for more details.

Journal

We would like to submit extended version of submitted papers for publication in a special issue of an international journal.