Investigations of the polar atmosphere with use of dynamical characteristics of the processes and self -organizing neural networks on ground-based multi-position optical observations

This article discusses questions connected with investigations of the polar atmosphere via ground-based multi-position optical observations. For this investigation a specially developed method is applied. It use the calculated dynamical parameters of the processes and self-organizing artificial neural networks. With self-organizing neural networks 5–7 classes have been extracted which can be associated with the regions having different physical properties. In particular, have been extracted regions intuitively coincident with polar stratospheric clouds and aurora.