Predicting the Astronomical Events in IraqBy Using Backpropagation and Radial Basis Function Networks: A Comparative Analysis | ||
| Al-Mustansiriyah Journal of Science | ||
| Article 1, Volume 23, Issue 7, October 2012, Pages 141-156 | ||
| Author | ||
| Ahmad Hashim Hussein Aal-Yhia | ||
| Abstract | ||
| This paper describes two neural networks Backpropagation network (BPN) and Radial basis function network (RBFN). Neural networks were applied for predicting of astronomical events: solar eclipse, lunar eclipse, sunrise and sunset for several Iraqi cities. Neural network toolbox in MATLAB was used for training and simulation each event network and the used data in the paper is real data. Each event network was implemented five times, and then the average was computed. The performances of BPN and RBFN are compared for each average in terms of the number of epochs, the taken time for training, the taken time for convergence and ratio of errors in results. The results show that RBFN is a more efficient and practical neural network than BPN. Keywords: radial basis function network, backpropagation network, astronomical events, prediction. | ||
|
Statistics Article View: 131 PDF Download: 89 |
||