Modelling stock markets forecasting using neural networks

This article is devoted to the substantiation of stock markets forecasting modelling using a neural network that describes the principles of the simulation algorithm im-plementation and the prospects for its application. The problems of traditional and classical forecasting systems, the theory of neural networks, the problems of improving the methods of analysis and improving the accuracy of stock market forecasts, simulating fuzzy models on the basis of sets of independent variables and the most informative factors of influence have been considered. The advantages of computational methods are analyzed for making up artificial neural networks simulating models that forecast exchange rates. Formulas for using the chosen forecasting method have been given as well as an explanation for the regression analysis. There exists an optimal combination for the assets and the most profitable investment period for each asset. The article emphasizes the growing rejection of the widely used classical economy and mathematical methods and models for adequate analysis and forecasting the development of financial and economic systems, which do not make it possible to effectively prevent significant and lastung crises at the stock markets. The scientific substantiation of the methodology for applying predictive modelling in choosing support system for fuzzy logic algorithms has been described. On the basis of a neural network system price dynamics forecasting at the stock market is simulated