|
Activity Number:
|
517
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Defense and National Security
|
| Abstract - #304400 |
|
Title:
|
Clustering Time Series Using Compression Dissimilarities and Wavelets
|
|
Author(s):
|
Patricia H. Carter*+
|
|
Companies:
|
Naval Surface Warfare Center
|
|
Address:
|
18444 Frontage Road, Dahlgren, VA, 22448,
|
|
Keywords:
|
time series ; clustering ; compression dissimilarity ; wavelet
|
|
Abstract:
|
Time series data mining tasks include the clustering and classification of time segments, the retrieval of segments similar to a given segment, and the summarization of the time series into epochs of similar behavior. Compression dissimilarities based on file and string compressors have been used to cluster and/or classify segments of time series. The underlying idea is that the more similar time series segment A and segment B are, the closer the size of the compressed AUB will be to the size of the compressed A. In this work we explore the utility of preprocessing the time series segments using various integer to integer wavelet transforms.
|
