Output details
13 - Electrical and Electronic Engineering, Metallurgy and Materials
University of Liverpool
Article title
Local prediction of non-linear time series using support vector regression
Type
D - Journal article
Title of journal
Pattern Recognition
Article number
-
Volume number
41
Issue number
5
First page of article
1539
ISSN of journal
00313203
Year of publication
2008
URL
-
Number of additional authors
1
Additional information
The paper presents an original method for multi-step head prediction, using delay co-ordinate embedding to transfer nonlinear time series to a reconstructed state space. The paper led to a National Grid project on UK national gas demand forecasting, and the project has resulted in a significant profit saving worth £130k ~£260k per day (Chris.aldridge@uk.ngrid.com), which has been submitted as an impact case for an associated UoA.
Interdisciplinary
-
Cross-referral requested
-
Research group
None
Proposed double-weighted
No
Double-weighted statement
-
Reserve for a double-weighted output
No
Non-English
No
English abstract
-