Content uploaded by Mohamed Medhat Gaber
Author content
All content in this area was uploaded by Mohamed Medhat Gaber on Mar 01, 2016
Content may be subject to copyright.
Knowledge Discovery from Sensor Data (SensorKDD)
Olufemi A. Omitaomu
Oak Ridge National Laboratory
1, Bethel Valley Road
Oak Ridge, TN 37831, USA
omitaomuoa@ornl.gov
Ranga Raju Vatsavai
Oak Ridge National Laboratory
1, Bethel Valley Road
Oak Ridge, TN 37831, USA
vatsavairr@ornl.gov
Auroop R. Ganguly
Oak Ridge National Laboratory
1, Bethel Valley Road
Oak Ridge, TN 37831, USA
gangulyar@ornl.gov
Nitesh V. Chawla
University of Notre Dame
Department of Computer Science and
Engineering, IN, USA
nchawla@cse.nd.edu
Joao Gama
University of Porto
Rua de Ceuta
Porto, Portugal
jgama@fep.up.pt
Mohamed Medhat Gaber
Monash University
900 Dandenong Road
Melbourne, Victoria, 3145, Australia
mohamed.gaber@infotech.mo
nash.edu.au
ABSTRACT
Wide-area sensor infrastructures, remote sensors, RFIDs, and
wireless sensor networks yield massive volumes of disparate,
dynamic, and geographically distributed data. As such sensors are
becoming ubiquitous, a set of broad requirements is beginning to
emerge across high-priority applications including adaptability to
climate change, electric grid monitoring, disaster preparedness
and management, national or homeland security, and the
management of critical infrastructures. The raw data from sensors
need to be efficiently managed and transformed to usable
information through data fusion, which in turn must be converted
to predictive insights via knowledge discovery, ultimately
facilitating automated or human-induced tactical decisions or
strategic policy based on decision sciences and decision support
systems.
Keeping in view the requirements of the emerging field of
knowledge discovery from sensor data, we took initiative to
develop a community of researchers with common interests and
scientific goals, which culminated into the organization of
SensorKDD series of workshops in conjunction with the
prestigious ACM SIGKDD International Conference of
Knowledge Discovery and Data Mining. In this report, we
summarize events at the Third ACM-SIGKDD International
Workshop on Knowledge Discovery form Sensor Data
(SensorKDD 2009).
1. INTRODUCTION
As wide-area sensor infrastructures, remote sensors, RFIDs, and
wireless sensor networks are becoming ubiquitous, the challenges
for the knowledge discovery community are expected to be
immense. On the one hand, dynamic data streams or events
require real-time analysis methodologies and systems, while on
the other hand centralized processing through high end computing
is also required for generating offline predictive insights, which in
turn can facilitate real-time analysis. The online and real-time
knowledge discovery imply immediate opportunities as well as
intriguing short- and long-term challenges for practitioners and
researchers in knowledge discovery. The opportunities would be
to develop new data mining approaches and adapt traditional and
emerging knowledge discovery methodologies to the requirements
of the emerging problems. In addition, emerging societal
problems require knowledge discovery solutions that are designed
to investigate anomalies, hotspots, changes, extremes and
nonlinear processes, and departures from the normal. The theme
for the 2009 SensorKDD workshop is around three inter-related
global priorities: climate change, energy assurance, and
infrastructural impacts. The workshop brings together researchers
from academia, government, and the industry working in various
aspects of knowledge discovery from sensor data.
1.1 Motivation
The expected ubiquity of sensors in the near future, combined
with the critical roles they are expected to play in high priority
application solutions, point to an era of unprecedented growth and
opportunities. The requirements described earlier imply
immediate opportunities as well as intriguing short- and long-term
challenges for practitioners and researchers in knowledge
discovery. In addition, the knowledge discovery and data mining
(KDD) community would be called upon, again and again, as
partners with domain experts to solve critical application solutions
in business and government, as well as in the domain sciences and
engineering.
The main motivation for this workshop stems from the increasing
need for a forum to exchange ideas and recent research results,
and to facilitate collaboration and dialog between academia,
government, and industrial stakeholders. Based on the positive
feedback from the previous workshop attendees and our own
experiences and interactions with the government agencies such
as the United States Department of Energy, United States
Department of Homeland Security, United States Department of
Defense, and involvement with numerous projects on knowledge
discovery from sensor data, we strongly believe in the
continuation of this workshop. We also believe that the ACM
SIGKDD conference is the right forum to organize this workshop
as it brings the KDD community together in this important area to
establish a much needed leadership position in research and
practice in the near term, as well as in the long term.
1.2 Previous SensorKDD Workshops
The previous two workshops – SensorKDD’07 and
SensorKDD’08 – held in conjunction with the 13
th
and 14
th
ACM
SIGKDD Explorations Volume 11, Issue 2 Page 84
SIGKDD Conference on Knowledge Discovery and Data Mining
respectively attracted several participants as well as many high
quality papers and presentations. The 2007 workshop was
attended by more than seventy registered participants. The
workshop program includes presentations by authors of six
accepted full papers and four invited speakers. The invited
speakers were Prof. Pedro Domingos of the University of
Washington, Prof. Joydeep Ghosh of the University of Texas,
Austin, Prof. Hillol Kargupta of the University of Maryland,
Baltimore County, and Dr. Brian Worley of the Oak Ridge
National Laboratory (ORNL). There were also poster
presentations by authors of six accepted short papers. The
extended versions of papers presented at the workshop were
developed into a book titled Knowledge Discovery from Sensor
Data [1], the first book published in this specific discipline.
The 2008 workshop was attended by more than 60 registered
participants. There were presentations by authors of seven
accepted full papers and six accepted short papers; the workshop
program also include presentations by two invited speakers –
Prof. Jiawei Han of the University of Illinois at Urbana-
Champaign and Dr. Kendra Moore of the Defense Advanced
Research Projects Agency. The extended versions of papers
presented at the 2008 workshop are scheduled for publication as
Springer's LNCS post-proceedings in 2009.
2. SUMMARY OF THE 2009 WORKSHOP
The highlights of the SensorKDD 2009 workshop are oral
presentations of accepted papers, presentations by invited
speakers, and oral presentations by the winners of the first
SensorKDD Cup, similar in spirit to the KDD Cup, but in line
with the workshop theme.
Based on a minimum of two reviews per paper, we selected six
full paper and eight short papers. In addition to the oral
presentations of accepted papers, the workshop featured three
invited speakers: Dr. Aurelie Lozano of IBM T.J. Watson
Research Center, Mr. Alessandro Donati of the European Space
Agency, Darmstadt, Germany, and Prof. Carlos Guestrin of
Carnegie Mellon University. For the SensorKDD Cup, we
challenged contestants with intriguing problems along
with geographically distributed and dynamic data from sensors
and model simulations. The datasets were provided and/or
compiled by leveraging the Oak Ridge National Laboratory
resources. Two winners were selected to present their results at
the workshop. All papers presented at the workshop are available
in [2].
2.1 Full Research Papers
Below is a list of accepted full papers and their respective authors:
• Handling Outliers and Concept Drift in Online Mass
Flow Prediction in CFB Boilers by J. Bakker, M.
Pechenizkiy, I. Zliobaite, A. Ivannikov, and T.
Karkkainen (Best Paper Award Winner).
• An Exploration of Climate Data Using Complex
Networks by Karsten Steinhaeuser, Nitesh V. Chawla,
and Auroop R. Ganguly (1
st
Best Student Paper Award
Winner).
• A Comparison of SNOTEL and AMSR-E Snow Water
Equivalent Datasets in Western U.S. Watersheds by
Cody L. Moser, Oubeidillah Aziz, Glenn A. Tootle,
Venkat Lakshmi, and Greg Kerr.
• EDISKCO: Energy Efficient Distributed in-Sensor-
Network K-center Clustering with Outliers by Marwan
Hassani, Emmanuel Muller, and Thomas Seidl.
• Phenological Event Detection from Multitemporal
Image Data by Ranga Raju Vatsavai.
• Mining in a Mobile Environment by Sean McRoskey,
James Notwell, Nitesh V. Chawla, and Christian
Poellabauer (2
nd
Best Student Paper Award Winner).
2.2 Short Research Papers
The following is a list of accepted short papers and their
respective authors:
• On the Identification of Intra-seasonal Changes in the
Indian Summer Monsoon by Shivam Tripathi and Rao
S. Govindaraju.
• Reduction of Ground-Based Sensor Sites for Spatio-
Temporal Analysis of Aerosols by Vladan
Radosavljevic, Slobodan Vucetic, and Zoran
Obradovic.
• OcVFDT: One-class Very Fast Decision Tree for One-
class Classification of Data Streams by Chen Li, Yang
Zhang, and Xue Li.
• A Frequent Pattern Based Framework for Event
Detection in Sensor Network Stream Data by Li Wan,
Jianxin Liao, and Xiaomin Zhu.
• Supervised Clustering via Principal Component
Analysis in a Retrieval Application by Esteban Garcia-
Cuesta, Ines M. Galvan, and Antonio J. de Castro.
• A Novel Measure for Validating Clustering Results
Applied to Road Traffic by Yosr Naija and Kaouther
Blibech Sinaoui.
• SkyTree: Scalable Skyline Computation for Sensor Data
by Jongwuk Lee and Seung-won Hwang.
• Clustering of Power Quality Event Data Collected via
Monitoring Systems Installed on the Electricity
Network by Mennan Guder, Nihan Kesim Cicekli,
Ozgul Salor, and Isik Cadirci.
2.3 SensorKDD’09 Cup Winners
The titles and authors of the SensorKDD’09 Cup winners are:
• Change Detection in Rainfall and Temperature Patterns
over India by Shivam Tripathi and Rao S. Govindaraju.
• Anomaly Detection and Spatio-Temporal Analysis of
Global Climate System by Mahashweta Das and
Srinivasan Parthasarathy.
3. CONCLUSIONS
Extracting knowledge and emerging patterns from sensor data is a
nontrivial task. The challenges for the knowledge discovery
community are expected to be immense. As evidenced from the
participation and quality of submissions to the previous three
SensorKDD workshops, it is clear that the “Knowledge Discovery
from Sensor Data or SensorKDD” is a growing area and an
important specialty (sub-area) within knowledge discovery. The
SIGKDD Explorations Volume 11, Issue 2 Page 85
SensorKDD workshop is proven to be an attractive forum for the
researchers from academia, industry and government, to exchange
ideas, initiate collaborations and lay foundation to the future of
this important and growing area. The workshop witnessed lively
participation from all quarters, generated interesting discussions
immediately after each presentation and as well as at the end of
the workshop. All participants agreed for continued patronage for
the SensorKDD workshop. In addition to the ACM workshop
proceedings, extended papers will be published as post workshop
proceedings in Springer's well-known “Lecture Notes in
Computer Science” series.
4. SPONSORSHIP
The SensorKDD’09 workshop was sponsored by the Geographic
Information Science and Technology (GIST) Group at Oak Ridge
National Laboratory, the Computational Sciences and Engineering
(CSE) Division at the Oak Ridge National Laboratory, and
Cooperating Objects Network of Excellence (CONET).
5. ACKNOWLEDGMENTS
We would like to thank our sponsors for their kind donations. We
would like to thank the authors of all submitted papers and
presenters. Their innovation and creativity has resulted in a strong
technical program. We are highly indebted to the program
committee members, whose reviews ensured the development of a
competitive and strong technical program. The program
committee listed in alphabetical order of last names are: Adedeji
B. Badiru, Eric Auriol, Albert Bifet, Michaela Black, Jose del
Campo-Avila, Andre Carvalho, Sanjay Chawla, Diane Cook,
Alfredo Cuzzocrea, Christie Ezeife, David J. Erickson III, Yi
Fang, Francisco Ferrer, James H. Garrett, Joydeep Ghosh, Bryan
L. Gorman, Sara Graves, Ray Hickey, Forrest Hoffman, Luke
(Jun) Huan, Volkan Isler, Vandana Janeja, Yu (Cathy) Jiao, Ralf
Klinkenberg, Miroslav Kubat, Vipin Kumar, Mark Last, Chang-
Tien Lu, Elaine Parros Machado de Sousa, Sameep Mehta,
Laurent Mignet, S. Muthu Muthukrishnan, George Ostrouchov,
Rahul Ramachandran, Pedro Rodrigues, Josep Roure, Bernhard
Seeger, Cyrus Shahabi, Shashi Shekhar, Lucio Soibelman,
Alexandre Sorokine, Eduardo J. Spinosa, Karsten Steinhaeuser,
Peng Xu, Eiko Yoneki, Philip S. Yu, Nithya Vijayakumar, and
Guangzhi Qu.
We would like to thank our invited speakers, Dr. Aurélie Lozano
of IBM T.J. Watson Research Center, Mr. Alessandro Donati of
European Space Agency, Darmstadt, Germany, and Prof. Carlos
Guestrin of
Carnegie Mellon University, who, despite their busy
schedules, readily agreed and delivered highly motivating and
informative talks. We would like to thank, Dr. Brian Worley,
Director, Computational Sciences and Engineering Division
(CSED), Oak Ridge National Laboratory (ORNL), for his
encouragement, support, and continued patronage of SensorKDD
workshop series, and Dr. Budhendra Bhaduri, Group Leader,
Geographic Information Science and Technology, CSED, ORNL,
for his enthusiastic support and sponsorship.
This workshop report was compiled by Dr. Olufemi A. Omitaomu
of the Computational Sciences and Engineering Division at Oak
Ridge National Laboratory. The workshop report has been co-
authored by employees of UT-Battelle, LLC, under contract DE-
AC05-00OR22725 with the U.S. Department of Energy. The
United States Government retains, and the publisher by accepting
the article for publication, acknowledges that the United States
Government retains, a non-exclusive, paid-up, irrevocable, world-
wide license to publish or reproduce the published form of this
manuscript, or allow others to do so, for United States
Government purposes.
6. REFERENCES
[1]
Auroop R. Ganguly, Joao Gama, Olufemi A. Omitaomu,
Mohamed M. Gaber, and Ranga Raju Vatsavai (2009).
Knowledge Discovery from Sensor Data. New York, NY:
CRC Press, January.
[2] Olufemi A. Omitaomu, Auroop R. Ganguly, Joao Gama,
Ranga Raju Vatsavai, Nitesh V. Chawla, and Mohamed
Medhat Gaber (2009). Proceedings of the Third
International Workshop on Knowledge Discovery from
Sensor Data. Paris, France: ACM Digital Library. Available
online:
http://portal.acm.org/toc.cfm?id=1601966&coll=GUIDE&dl
=GUIDE&type=proceeding&idx=SERIES939&part=series&
WantType=Proceedings&title=KDD&CFID=49782169&CF
TOKEN=54456709.
About the Workshop Organizers:
Dr. Olufemi A. Omitaomu is a research scientist in the
Computational Sciences and Engineering Division at the Oak
Ridge National Laboratory. He is also an adjunct assistant
professor at the University of Tennessee, Knoxville. His research
interests include streaming and real-time data mining,
infrastructure modeling and analysis, machine learning, and signal
processing. He received Ph.D. in information engineering from
the University of Tennessee. He has published in top peer-
reviewed journals and conferences; co-organized and co-chaired
workshop and sessions at professional conferences including the
ACM Workshop on Knowledge Discovery from Sensor Data held
in conjunction with ACM SIGKDD 2007 and ACM SIGKDD
2008. He previously worked as a data analyst with Mobil
Exploration and Production Company for more than five years.
Dr. Auroop R. Ganguly is a research scientist within the
Computational Sciences and Engineering division of the Oak
Ridge National Laboratory since 2004. His research interests are
climate change impacts, geoscience informatics, civil and
environmental engineering, computational data sciences, and
knowledge discovery. Prior to ORNL, he has more than five years
of experience in the software industry, specifically Oracle
Corporation and a best-of-breed company subsequently acquired
by Oracle, and about a year in academia, specifically at the
University of South Florida in Tampa. He has a PhD from the
Civil and Environmental Engineering department of the
Massachusetts Institute of Technology, several years of research
experience with a group at the MIT Sloan School of Management,
experience in private consulting, and a wide range of peer-
reviewed publications spanning multiple disciplines. Currently, he
is also an adjunct professor at the University of Tennessee in
Knoxville.
SIGKDD Explorations Volume 11, Issue 2 Page 86
Dr. Joao Gama is a researcher at LIAAD-INESC Porto LA, the
Laboratory of Artificial Intelligence and Decision Support of the
University of Porto. His main research interest is learning from
Data Streams. He has published several articles in change
detection, learning decision trees from data streams, hierarchical
clustering from streams, among others. Editor of special issues on
Data Streams in Intelligent Data Analysis, Journal of Universal
Computer Science, and New Generation Computing. Co-chair of
ECML 2005 Porto, Portugal 2005, Conference chair of Discovery
Science 2009, and of a series of Workshops on Knowledge
Discovery in Data Streams, ECML 2004, Pisa, Italy, ECML 2005,
Porto, Portugal, ICML 2006, Pittsburg, US, ECML 2006 Berlin,
Germany, and SAC2007, Korea. Together with Dr. Mohamed M.
Gaber edited the book Learning from Data Streams Processing
Techniques in Sensor Networks, published by Springer.
Dr. Ranga Raju Vatsavai has been conducting research in the area
of spatiotemporal databases and data mining for the past 15 years.
Before joining the Oak Ridge National Laboratory (ORNL) as a
research scientist, he worked at IBM-Research (2004-06; IIT-
Delhi campus), U of Minnesota (1999-2004; Twin-cities campus,
MN), AT&T Labs (1998; Middletown, NJ), Center for
Development of Advanced Computing (1995-98; C-DAC, U of
Pune campus, India), and National Forest Data Management
Center (1990-95; FRI Campus, Dehradun, India). He has
published over thirty peer-reviewed articles and served on
program committees of several international conferences (KDD,
ICTAI, SSTDM). He was also involved in the design and
development of several highly successful software systems
(UMN-MapServer - a world leading open source WebGIS,
*Miner - a spatiotemporal data mining workbench, EASI/PACE
classification modules, and first parallel softcopy photogrammetry
system for IRS-1C/1D satellites). His broad research interests are
centered on spatial, spatiotemporal databases and data mining,
and computational geoinformatics; in particular he is interested in
statistical pattern recognition, semi-supervised learning, multiple
classifier systems, time series analysis and forecasting,
information retrieval, uncertainty and error handling.
Dr. Nitesh V. Chawla is an assistant professor at the University of
Notre Dame. Dr. Chawla's research interests are broadly in the
areas of data mining, machine learning, pattern recognition, and
their applications. More specifically his research has focused on
learning from massive datasets, distributed data mining/machine
learning, ensemble techniques, cost/distribution sensitive learning,
feature selection, and semi-supervised learning. His research has
also focused on the inter-disciplinary applications such as
intelligent scientific visualization, biometrics, bioinformatics,
natural language processing, and customer analytics.
Dr. Mohamed Medhat Gaber is a research Fellow at Monash
University, Australia. He has published more than 60 papers.
Mohamed is the co-editor of the book: Learning from Data
Streams: Processing Techniques in Sensor Networks, published
by Springer in 2007. His research interests include data stream
mining, wireless sensor networks and context-aware computing.
Mohamed has served in the program committees of several
international and local conferences and workshops in the area of
data mining and context-aware computing. He was the co-chair of
the IEEE International Workshop on Mining Evolving and
Streaming Data held in conjunction with ICDM 2006,
International Workshop on Knowledge Discovery from
Ubiquitous Data Streams held in conjunction with ECML/PKDD
2007, and the First and Second International Workshop on
Knowledge Discovery from Sensor Data held in conjunction with
ACM SIGKDD 2007/2008.
SIGKDD Explorations Volume 11, Issue 2 Page 87
![Fig. 1. Real-world examples of how human behavior directly influences sensor data time patterns: highway traffic volume [7], home water demand [8], indoor residential sound level [9] and electricity demand [10]. Not only the evolution is cyclic but also highly correlated among different daily-life dimensions.](https://www.researchgate.net/profile/Pedro-Rodrigues-48/publication/221653930/figure/fig1/AS:305565449375759@1449863888295/Real-world-examples-of-how-human-behavior-directly-influences-sensor-data-time-patterns_Q320.jpg)
