Symbolic Time Series Analysis (STSA) for Anomaly Detection
Devendra K. Tolani
Intelligent Automation, Inc.
Tuesday, December 6, 2005 15:00-16:00,
NIST North (820), Room 145
Gaithersburg
Tuesday, December 6, 2005 13:00-14:00,
Room 4550
Boulder
Abstract:
This presentation will introduce a novel concept of anomaly detection in complex dynamical systems using tools of symbolic time series analysis (STSA),
finite state automata, and pattern recognition,
where time series data of the observed variables on the fast time-scale are analyzed at slow time-scale epochs for early detection of (possible) anomalies.
The anomaly detection problem is separated into two parts:
(i) forward (or analysis) problem of pattern discovery to identify variations in the anomalous behavior patterns, compared to the nominal behavior;
and (ii) inverse (or synthesis) problem of pattern recognition to infer parametric or nonparametric changes (i.e., anomalies)
based on the learnt patterns and observed stationary response.
The performance of the proposed method will be compared to existing pattern recognition tools, such as neural networks and principal component analysis,
for detection of small changes in the statistical characteristics of the observed data sequences.
These concepts will be explained in detail by the results from a number of experimental and simulation studies.
Speaker Bio:
Devendra Kumar Tolani received his B-Tech (Honors) in Mechanical Engineering from the Indian Institute of Technology, Kharagpur, India, in 1999.
Before joining Pennsylvania State University for graduate studies, he worked as an Engineer at Tata Engineering.
He has two MS degrees, one in Mechanical and the other in Electrical Engineering, both from Penn State.
He received his PhD in Mechanical Engineering from Penn State in 2005.
The topic of his dissertation was Integrated Health Management and Control of Complex Dynamical Systems.
His general research interests include: Control Theory, Signal Processing and Analysis, and Discrete Event Systems.
His specific areas of interest include Diagnostics Prognostics and Health Management (DPHM), C4ISR, and Data Driven Modeling.
He is currently working as a Research Scientist at Intelligent Automation, Inc.
Presentation Slides: PDF
Contact: F. HuntNote: Visitors from outside NIST must contact
Robin Bickel; (301) 975-3668;
at least 24 hours in advance.
|
