Neural net forecasting for geomagnetic activity

Hernández, José Falcón; Tajima, T.; Horton, W.

doi:10.1029/93gl02848

Cited by 71 publications

(46 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nonlinear models for the AL prediction, driven by the solar wind and IMF parameters, have been pioneered by Bargatze et al (1985) and further developed by Vassiliadis (1994), Price et al (1994) and . Neural networks have been introduced into prediction of auroral indices along with nonlinear models (ARMA and MA filters) by Hernandez et al (1993) and later on by Gleisner and Lundstedt (1997) and Weigel et al (1999). Following the steps of Hernandez et al (1993), our model uses neural nets to construct nonlinear models using Autoregressive Moving Average with eXogenous input(s) (ARMAX).…”

Section: Introductionmentioning

confidence: 99%

“…Neural networks have been introduced into prediction of auroral indices along with nonlinear models (ARMA and MA filters) by Hernandez et al (1993) and later on by Gleisner and Lundstedt (1997) and Weigel et al (1999). Following the steps of Hernandez et al (1993), our model uses neural nets to construct nonlinear models using Autoregressive Moving Average with eXogenous input(s) (ARMAX). The data and the challenges of the data sets in correlation with the model are discussed in Sect.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On the prediction of the auroral westward electrojet index

Amariutei

Ganushkina

2012

Ann. Geophys.

View full text Add to dashboard Cite

Abstract. An ARMAX based model, to forecast the evolution of the of AL index, is developed. The model has been trained and validated using neural networks with the half wave rectifier (V B s ) as input. It is shown that the model posses a good, reliable forecasting ability, including periods of intense geomagnetic activity. Prediction efficiency of the model is discussed in the context of 1 min resolution output smoothed over 7 min.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

On the prediction of the auroral westward electrojet index

Amariutei

Ganushkina

2012

Ann. Geophys.

View full text Add to dashboard Cite

show abstract

“…The dimension analyses of AE time series gave evidence of low effective dimension in the system (Vassiliadis et al, 1990;Sharma et al, 1993). Further elaboration of this hypothesis resulted in creating space weather forecasting tools based on local-linear filters (Price et al, 1994;Vassiliadis et al, 1995;Valdivia et al, 1996), data-derived analogues (Klimas et al, 1997;Horton et al, 1999), and neural networks (Hernandes et al, 1993;Gleisner and Lundstedt, 1997;Weigel et al, 2002). The low dimensional organized behavior of the magnetosphere on global scales is also evident in many in situ observations of the large-scale features of substorms by many spacecrafts including INTERBALL and GEOTAIL missions (Petrukovich et al, 1998;Nagai et al, 1998;Ieda et al, 1998).…”

Section: Introductionmentioning

confidence: 93%

Combining global and multi-scale features in a description of the solar wind-magnetosphere coupling

et al. 2003

View full text Add to dashboard Cite

Abstract. The solar wind-magnetosphere coupling during substorms exhibits dynamical features in a wide range of spatial and temporal scales. The goal of our work is to combine the global and multi-scale description of magnetospheric dynamics in a unified data-derived model. For this purpose we use deterministic methods of nonlinear dynamics, together with a probabilistic approach of statistical physics. In this paper we discuss the mathematical aspects of such a combined analysis. In particular we introduce a new method of embedding analysis based on the notion of a mean-field dimension. For a given level of averaging in the system the mean-filed dimension determines the minimum dimension of the embedding space in which the averaged dynamical system approximates the actual dynamics with the given accuracy. This new technique is first tested on a number of wellknown autonomous and open dynamical systems with and without noise contamination. Then, the dimension analysis is carried out for the correlated solar wind-magnetosphere database using vB S time series as the input and AL index as the output of the system. It is found that the minimum embedding dimension of vB S − AL time series is a function of the level of ensemble averaging and the specified accuracy of the method. To extract the global component from the observed time series the ensemble averaging is carried out over the range of scales populated by a high dimensional multiscale constituent. The wider the range of scales which are smoothed away, the smaller the mean-field dimension of the system. The method also yields a probability density function in the reconstructed phase space which provides the basis for the probabilistic modeling of the multi-scale dynamical features, and is also used to visualize the global portion of the solar wind-magnetosphere coupling. The structure of its input-output phase portrait reveals the existence of two energy levels in the system with non-equilibrium dynamical features such as hysteresis which are typical for non-equilibrium phase transitions. Further improvements in space weather forecasting tools may be achieved by a combiCorrespondence to: A. Y. Ukhorskiy (sasha@astro.umd.edu) nation of the dynamical description for the global component and a statistical approach for the multi-scale component.

show abstract

“…Therefore, the algorithms make it possible to construct a forecast model even though the relationship between cause and effect is not clearly understood. Neural networks have been broadly used in space weather applications such as the prediction of solar activities (Wang 2000;Gong et al 2004;Wang et al 2008;Qahwaji et al 2007Qahwaji et al , 2008Henwood et al 2010) and geomagnetic activities (Lundstedt 1992;Hernandez et al 1993;Freeman et al 1993;Valach et al 2009;Ji et al 2013). SPE prediction models have been developed by neu-…”

Section: Introductionmentioning

confidence: 99%

Forecast of Solar Proton Events With Noaa Scales Based on Solar X-Ray Flare Data Using Neural Network

Jeong¹,

Lee²,

Moon³

et al. 2014

Journal of The Korean Astronomical Society

View full text Add to dashboard Cite

Abstract:In this study we develop a set of solar proton event (SPE) forecast models with NOAA scales by Multi Layer Perceptron (MLP), one of neural network methods, using GOES solar X-ray flare data from 1976 to 2011. Our MLP models are the first attempt to forecast the SPE scales by the neural network method. The combinations of X-ray flare class, impulsive time, and location are used for input data. For this study we make a number of trials by changing the number of layers and nodes as well as combinations of the input data. To find the best model, we use the summation of F-scores weighted by SPE scales, where F-score is the harmonic mean of PODy (recall) and precision (positive predictive value), in order to minimize both misses and false alarms. We find that the MLP models are much better than the multiple linear regression model and one layer MLP model gives the best result.

show abstract

Neural net forecasting for geomagnetic activity

Cited by 71 publications

References 19 publications

On the prediction of the auroral westward electrojet index

On the prediction of the auroral westward electrojet index

Combining global and multi-scale features in a description of the solar wind-magnetosphere coupling

Forecast of Solar Proton Events With Noaa Scales Based on Solar X-Ray Flare Data Using Neural Network

Contact Info

Product

Resources

About