Output details
11 - Computer Science and Informatics
University of East Anglia
Article title
A run length transformation for discriminating between auto regressive time series
Type
D - Journal article
Title of journal
Journal of Classification
Article number
-
Volume number
n/a
Issue number
-
First page of article
n/a
ISSN of journal
1432-1343
Year of publication
2013
URL
-
Number of additional authors
1
Additional information
<15>Model-based approaches are a popular method for measuring time series similarity in a range of disciplines such as speech processing and econometrics. The problem with model-based approaches is that the model fitting and updating can be time consuming for big data. We show both theoretically and experimentally that the run length distribution can asymptotically approximate the autoregressive model and yet can be fitted in linear time and updated in constant time. This result will have impact on problems where the sheer volume of streaming data makes refitting the model unfeasible for a timely classification.
Interdisciplinary
-
Cross-referral requested
-
Research group
None
Citation count
0
Proposed double-weighted
No
Double-weighted statement
-
Reserve for a double-weighted output
No
Non-English
No
English abstract
-
