Abstract
Nowadays there exists a large variety of forecasting methods based on neural networks technologies which allow to adequately simulate the nonlinear processes with noisy data in slowly varying markets. However, in the conditions of strong turbulence they are not capable to promptly response to swiftly changing environment, and, hence, to market conjuncture changes. It has grown in a very serious problem, as the decisions accepted on the basis of these technologies can be resulted in movement of hundreds billion dollars in the financial markets, and any incorrectly revealed tendency can lead to large losses.We will notice that this problem has recently arisen due to joining of new world players to the common economics, and due to, as a consequence, decrease in controllability of the international system. In the presented work the dynamic approach will be considered which properly operates in swiftly changing markets. It is based on a combination of neural networks technologies with algorithms of quantum mechanics calculations.