Skip to Main Content
Frequently Asked Questions
Submit an ETD
Global Search Box
Need Help?
Keyword Search
Participating Institutions
Advanced Search
School Logo
Files
File List
osu1261540196.pdf (23.56 MB)
ETD Abstract Container
Abstract Header
Spatio-Temporal Anomaly Detection
Author Info
Das, Mahashweta
Permalink:
http://rave.ohiolink.edu/etdc/view?acc_num=osu1261540196
Abstract Details
Year and Degree
2009, Master of Science, Ohio State University, Computer Science and Engineering.
Abstract
Recent advances in computational sciences have led to the generation and utilization of enormous amounts of spatio-temporal data in numerous scientific disciplines, such as wireless sensor networks, bioinformatics, astrophysics and computational fluid dynamics. The need to efficiently handle and effectively analyze such humongous amount of data has led to the application of data mining techniques in this domain of research. Spatio-temporal data mining is the process of maneuvering such datasets to extract interesting knowledge and meaningful insights from the data. In this thesis, we propose a spatio-temporal and a temporal-spatial anomaly detection algorithm in the context of wireless sensor network application. The latter can also be interpreted as a prediction model for spatio-temporal datasets. The anomaly detection algorithms identify local abnormalities in sensor data collected over space and time, independent of the global analytical view presented by the entire dataset. The first of our proposed algorithms identifies candidate spatial outliers by computing an outlierness indicator, which we call ANOI (Antimonotonic Outlierness Indicator) and then determines the spatio-temporal outliers by checking the stability of the candidates over time. Our novel idea of antimonotonicity in an outlier detection framework helps us to prune the search space and reduce computational complexity, which is a serious bottleneck for neighborhood-based outlier detection methodologies. The second algorithm first models time series data and then refines the temporal estimates by integrating spatial association information. Since the model predicts short-term energy availability based on past historical records, it can also assist wireless sensor nodes to automatically adapt to changing environmental conditions and use its energy harvesting and activity scheduling abilities intelligently. Experimental results on climate data empirically demonstrate the effectiveness of both the approaches.
Committee
Srinivasan Parthasarathy, PhD (Advisor)
Gagan Agrawal, PhD (Committee Member)
Pages
74 p.
Subject Headings
Computer Science
Keywords
Anomaly Detection
;
Spatio-Temporal Mining
Recommended Citations
Refworks
EndNote
RIS
Mendeley
Citations
Das, M. (2009).
Spatio-Temporal Anomaly Detection
[Master's thesis, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1261540196
APA Style (7th edition)
Das, Mahashweta.
Spatio-Temporal Anomaly Detection.
2009. Ohio State University, Master's thesis.
OhioLINK Electronic Theses and Dissertations Center
, http://rave.ohiolink.edu/etdc/view?acc_num=osu1261540196.
MLA Style (8th edition)
Das, Mahashweta. "Spatio-Temporal Anomaly Detection." Master's thesis, Ohio State University, 2009. http://rave.ohiolink.edu/etdc/view?acc_num=osu1261540196
Chicago Manual of Style (17th edition)
Abstract Footer
Document number:
osu1261540196
Download Count:
1,205
Copyright Info
© 2009, all rights reserved.
This open access ETD is published by The Ohio State University and OhioLINK.