%0 Journal Article
%A M. R. Ghaffari Razin,
%A A. Mohammadzadeh,
%T Regional Ionosphere Modeling Using Artificial Neural Networks and Polynomial Fitting Over Iran
%J Journal of Geomatics Science and Technology
%V 4
%N 3
%U http://jgst.issge.ir/article-1-280-en.html
%R
%D 2015
%K Neural Networks, Total Electron Content, Perceptron Model, Back propagation algorithm, Polynomial fitting,
%X In this paper, 3-layer perceptron Neural Network has been used with 5 neuron in hidden layer for modeling the Ionospheric Total Electron Content (TEC) Over Iran. For this purpose, 25 GPS station from IPGN is used. These 25 stations are located within a range of approximately 24oN to 40oN and 44oE to 64oE. Evaluation of the results has been applied with 1 GPS station in Tehran. The station is equipped with ionosonde. So it is possible to calculate independently the TEC at the station. Minimum relative error obtained from evaluation is 0.73% and maximum relative error is 34.66 %. In this research, for the evaluation of artificial neural networks in estimating the TEC, a polynomial of degree 3 with 11 coefficients are used. Comparison of the relative error from polynomial model and relative error from neural network, illustrate the superiority of the neural model with respect to polynomial in this region. The number of neurons in hidden layer of neural network and the order and coefficients of the polynomial used in this paper is determined by trial and error, and by taking the minimum relative error for the results.
%> http://jgst.issge.ir/article-1-280-en.pdf
%P 51-60
%& 51
%!
%9 Research
%L A-10-25-65
%+
%G eng
%@ 2322-102X
%[ 2015