DocumentCode :
3517619
Title :
Sequential Modeling of D_st Dynamics with SEEk Trained Recurrent Neural Networks
Author :
Ouarbya, Lahcen ; Mirikitani, Derrick Takeshi
Author_Institution :
Dept. of Comput., Univ. of London, London, UK
fYear :
2010
fDate :
27-29 Jan. 2010
Firstpage :
32
Lastpage :
37
Abstract :
A sequential framework for modeling magnetospheric plasma interactions with a SEEK trained recurrent neural network is proposed. An overview of the state-space modeling framework is provided, along with a review of previous Kalman trained neural models. The proposed algorithm is described and is evaluated against an EKF trained RNN and a gradient based model. The exogenous inputs to the RNNs consist of three parameters, bz, b2, and by 2, where b, bz, and by represent the magnitude, the southward and azimuthal components of the interplanetary magnetic field (IMF) respectively. It was found that the SEEK trained recurrent neural network outperforms other neural time series models trained with the extended Kalman filter, and gradient descent learning. The numerical simulations suggest that the SEEK filter provides superior tracking capabilities than the EKF, resulting in accurate forecast of the Dst index.
Keywords :
Kalman filters; gradient methods; interplanetary magnetic fields; learning (artificial intelligence); magnetosphere; nonlinear filters; physics computing; plasma interactions; plasma simulation; recurrent neural nets; state-space methods; time series; Dst index forecasting; Kalman trained neural models; SEEk trained recurrent neural networks; extended Kalman filter; gradient based model; gradient descent learning; interplanetary magnetic field; magnetospheric plasma interaction modelling; neural time series models; numerical simulations; sequential modeling; state-space modeling framework; Bayesian methods; Earth; Inference algorithms; Kalman filters; Magnetic fields; Magnetic separation; Magnetosphere; Plasma simulation; Recurrent neural networks; Storms; Recurrent Neural Networks; SEEK filter;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Systems, Modelling and Simulation (ISMS), 2010 International Conference on
Conference_Location :
Liverpool
Print_ISBN :
978-1-4244-5984-1
Type :
conf
DOI :
10.1109/ISMS.2010.17
Filename :
5416125
Link To Document :
بازگشت