Jade, A. M. ; Jayaraman, V. K. ; Kulkarni, B. D. (2006) Improved time series prediction with a new method for selection of model parameters Journal of Physics A: Mathematical and General, 39 (30). L483-L491. ISSN 0305-4470
![]()
|
PDF
- Publisher Version
127kB |
Official URL: http://iopscience.iop.org/0305-4470/39/30/L01
Related URL: http://dx.doi.org/10.1088/0305-4470/39/30/L01
Abstract
A new method for model selection in prediction of time series is proposed. Apart from the conventional criterion of minimizing RMS error, the method also minimizes the error on the distribution of singularities, evaluated through the local Holder estimates and its probability density spectrum. Predictions of two simulated and one real time series have been done using kernel principal component regression (KPCR) and model parameters of KPCR have been selected employing the proposed as well as the conventional method. Results obtained demonstrate that the proposed method takes into account the sharp changes in a time series and improves the generalization capability of the KPCR model for better prediction of the unseen test data.
Item Type: | Article |
---|---|
Source: | Copyright of this article belongs to Institute of Physics. |
ID Code: | 17426 |
Deposited On: | 16 Nov 2010 08:40 |
Last Modified: | 17 May 2016 02:03 |
Repository Staff Only: item control page