An automatic method for estimating the geomagnetic field

Abstract

We introduce a new method for estimating the geomagnetic field. The method is based on a combination of a wavelet transform with radial basis neural networks. In the method, the recorded geomagnetic field variations are decomposed into different-scale components and the degree of disturbance of each component is estimated, enabling the conclusion about the field state. For the verification of the method, we used geomagnetic data from the "Paratunka" station (Paratunka, Kamchatka region, data registration is carried out by IKIR FEB RAS). Analysis of the spectral-temporal characteristics of geomagnetic field variations during periods of moderate and strong magnetic storms was performed. Weak perturbations were detected in the geomagnetic field before the storms. The obtained results have confirmed the effectiveness of the proposed method.

A neural network system for estimation of the geomagnetic field disturbance

“…Using multiresolution wavelet decomposition [13], we obtain the coefficients of the detailing components for the scales j = −1, −2, ..., −6 and calculate their absolute values. Obtained coefficients characterize disturbance of the geomagnetic field and in periods of high geomagnetic activity, their absolute values increase significantly [14].…”

“…The output layers of neural networks determine the probability of belonging of input images to the appropriate class. The process of construction and training of neural networks is described in detail in the paper [14].…”

Method of analysis of geomagnetic data based on wavelet transform and threshold functions

“…Based on the neural networks, the authors [25] suggested an algorithm for interplanetary magnetic field data processing and Dst-index calculation. The authors of that paper suggest an approach based on the combination of neural networks with wavelet transform and show the efficiency of joint application of mathematical apparatus data in comparison with a neural network in the problems of analysis of natural time series with complicated structures [26][27][28][29], in particular, for geomagnetic data analysis [28,29]. It is shown in these papers that wavelet transform allows us to investigate the data structure in detail and to detect informative components which, in their turn, improve the procedure of neural network training and its performance efficiency.…”

Interactive system for geomagnetic data analysis

“…Variations of the Earth's magnetic field contain important information on the processes in the magnetosphere that occur during the period of high solar activity. Currently, methods [3][4][5][6] and application software for the processing and analysis of geophysical data are being developed. They provide the users with convenient tools for conducting experimental and theoretical studies (eg.…”

“…The authors of the works [2,[9][10][11], on the basis of DataMining applications, developed methods for automating the work of experts (creating so-called "electronic experts") to solve the problems of primary data processing, to detect industrial and physical anomalies and to form world databases. Moreover, an algorithm based on the application of continuous wavelet transform [12] and threshold functions, first proposed in [3], showed high efficiency in detection of industrial noise.…”