A. I. Loskutov, S. V. Sirota, A. N. Sakulin

Given article is devoted developed on built in language Matlab 2010a to the automated information system of forecasting financial time series with use of hybrid neural networks, subsequently, received name AIS «NADEZHDA». The basic problems of forecasting of financial time series are considered, and also the decision of a problem of forecasting of the price of closing of share price of Open Society «Gazprom» on four last trading days is considered. Preparation of training sequences on an example of stock quotes of Open Society «Gazprom» is considered. Test tests of the given system where as training and testing sequences stock quotes of Open Society «Gazprom» acted are conducted and results of forecasting are shown by the given system. The result of an experimental research has shown high results of forecasting financial time series (stock quote of Open Society «Gazprom») with the minimum error of the forecast that is the important factor for the further development of the given system.