Artificial intelligence

Abstract: In the paper a forecasting method using a parallel-hierarchical (PH) network was developed. It uses the idea of smoothing the empirical data with the hyperbole. It is shown that this method uses the values of previous time series, the smoothing with the hyperbole and PH network data for prediction. The developed method is more efficient for real time systems due to the use of PH network combined with smoothing with a hyperbola; with the implementation of prediction operations of energy center laser beam image position for optical communication systems compared to conventional neural networks.

Keywords: parallel-hierarchical (PH) network, the images of laser beam spots, the statistical observation, time series, prediction

References:

1. Eliseev I.I. International statistics / I.I. Eliseev, Т.V. Kosteeva, L.М. Homenko. – Minsk : High. school, 1995. – P. 58-63.
2. Nafez Shweiki Method for Training of a Parallel-Hierarchical Network, Based on Population Coding for Processing of Extended Laser Paths Images / [Leonid I. Timchenko, Yuriy F. Kutaev, Volodymyr P. Kozhemyako et all.] // Proceedings of SPIE. – 2002. – Volume 4790. – P. 465-479.
3. Timchenko L.I. Convergent and divergent processes in real and artificial neural networks // Vestnik VPI. – 1997. – No 1. – P. 5-10.
4. ChetirkinЕ.М.Statisticalmethodsofpredicting.–M.:Statistics,1987–Р.200.
5. Segmentation of gray scale images based on connectivity : [мonograph] / [А.L.Zeleznyak, L.I. Timchenko, О.І. Stasyuk et all.]. - K.:SETUT, 2008 – p.144.
6. Timchenko L.I. Multi-stage parallel-hierarchical network as a model of neural computational scheme // Cybernetics and Systems Analysis. – 2000. – No 2. – P. 114-134.
7. The use of parallel-hierarchical method of laser beams spots image recognition. – International scientific and technical conference «Multiprocessor computing and control systems – 2009» / L.I. Timchenko, V.V. Melnikov, N. I. Kokriatskaia. – c. Taganrog. – P. 147-150.
8. The method of finding the energy centers laser paths fragments in real-time for pattern recognition. Proceedings of the 9th All-Ukrainian International Conference «Signal and Image Processing and Pattern Recognition» / L.I. Timchenko, V.V.Melnikov, N. I. Kokriatskaia. – c. Kiev. – 2008. – P. 213-216.
9. Thetheotyofstatistics/VashkovP.G.,PasterP.І.,StorozukV.P.–K.:Lubid,2001.–P.8-9.
10. Kopren U. The sampling methods. – Moscow : Statistics, 1976. 11.Method of reference tunnel formation for improving forecast results of the laser beams spot images behavior / [L. Timchenko, A. Poplavskyy, N. Petrovskyi,J.] // Opt. Eng. – 2011. – No 50, 117007. – DOI:10.1117/1.3655502. 12Borovykov V.P. The art of computer data analysis. – St. Petersburg. : Peter, 2003. – P. 688.