Advanced

Auroral electrojet predictions with dynamic neural networks

Gleisner, Hans and Lundstedt, Henrik LU (2001) In Journal of Geophysical Research 106(A11). p.24541-24550
Abstract
Neural networks with internal feedback from the hidden nodes to theinput [Elman, 1990 are developed for prediction of the auroralelectrojet index AE from solar wind data. Unlike linear and nonlinearautoregressive moving-average (ARMA) models, such networks are free todevelop their own internal representation of the recurrent statevariables. Further, they do not incorporate an explicit memory for paststates; the memory is implicitly given by the feedback structure of thenetworks. It is shown that an Elman recurrent network can predict around70 of the observed AE variance using a single sample of solar winddensity, velocity, and magnetic field as input. A neural network withidentical solar wind input, but without a feedback mechanism,... (More)
Neural networks with internal feedback from the hidden nodes to theinput [Elman, 1990 are developed for prediction of the auroralelectrojet index AE from solar wind data. Unlike linear and nonlinearautoregressive moving-average (ARMA) models, such networks are free todevelop their own internal representation of the recurrent statevariables. Further, they do not incorporate an explicit memory for paststates; the memory is implicitly given by the feedback structure of thenetworks. It is shown that an Elman recurrent network can predict around70 of the observed AE variance using a single sample of solar winddensity, velocity, and magnetic field as input. A neural network withidentical solar wind input, but without a feedback mechanism, onlypredicts around 45 of the AE variance. It is also shown that fourrecurrent state variables are optimal: the use of more than four hiddennodes does not improve the predictions, but with less than that theprediction accuracy drops. This provides an indication that theglobal-scale auroral electrojet dynamics can be characterized by a smallnumber of degrees of freedom. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Ionosphere: Current systems, Ionosphere: Modeling and forecasting, Magnetospheric Physics: Solar wind/magnetosphere interactions
in
Journal of Geophysical Research
volume
106
issue
A11
pages
24541 - 24550
publisher
American Geophysical Union
external identifiers
  • Scopus:39449102066
ISSN
2156-2202
DOI
10.1029/2001JA900046
language
English
LU publication?
yes
id
4067db84-ac43-4ba1-b908-173b2a620c49 (old id 130125)
date added to LUP
2007-07-13 10:05:18
date last changed
2016-10-13 04:45:01
@misc{4067db84-ac43-4ba1-b908-173b2a620c49,
  abstract     = {Neural networks with internal feedback from the hidden nodes to theinput [Elman, 1990 are developed for prediction of the auroralelectrojet index AE from solar wind data. Unlike linear and nonlinearautoregressive moving-average (ARMA) models, such networks are free todevelop their own internal representation of the recurrent statevariables. Further, they do not incorporate an explicit memory for paststates; the memory is implicitly given by the feedback structure of thenetworks. It is shown that an Elman recurrent network can predict around70 of the observed AE variance using a single sample of solar winddensity, velocity, and magnetic field as input. A neural network withidentical solar wind input, but without a feedback mechanism, onlypredicts around 45 of the AE variance. It is also shown that fourrecurrent state variables are optimal: the use of more than four hiddennodes does not improve the predictions, but with less than that theprediction accuracy drops. This provides an indication that theglobal-scale auroral electrojet dynamics can be characterized by a smallnumber of degrees of freedom.},
  author       = {Gleisner, Hans and Lundstedt, Henrik},
  issn         = {2156-2202},
  keyword      = {Ionosphere: Current systems,Ionosphere: Modeling and forecasting,Magnetospheric Physics: Solar wind/magnetosphere interactions},
  language     = {eng},
  number       = {A11},
  pages        = {24541--24550},
  publisher    = {ARRAY(0xaab7f68)},
  series       = {Journal of Geophysical Research},
  title        = {Auroral electrojet predictions with dynamic neural networks},
  url          = {http://dx.doi.org/10.1029/2001JA900046},
  volume       = {106},
  year         = {2001},
}