IDENTIFICATION METHODS OF NATURAL SIGNAL STRUCTURES

  1. Institute of Cosmophysical Research and Radio Wave Propagation of the Far-Eastern Branch of Russian Academy of Science

In the report methods of identification of regular components and specific features of difficult natural signals
are suggested. The proposed methods are significant in geophysics where there is a problem of identification of
anomalous effects which have important information on objects from which these signals were received. Identi-
fication of regular components was carried out by neural networks; the method of network training and the way
of formation of training sample are offered. Specific features define detailing components of signal and anoma-
lous structure, their identification is based on threshold methods applying wavelet-technology. The proposed
theoretical device was tested on real geophysical data registered at the Institute of Cosmophysical Research and
Radio Wave Propagation of the Far-Eastern Branch of Russian Academy of Science. The obtained theoretical
and experimental results of the investigation confirmed its efficiency in comparison with existing analogues.
Using this instrument, a system of analysis and forecast of the behavior of geophysical signals important char-
acteristics was developed.