The forecastings of future changes in the trend's direction of stock quotes by neural networks and fuzzy systems

The decision support system for a stock trader based on neural networks and fuzzy logic is proposed. An application of the fractal analysis and the shapes recognition on charts of the financial time series gives an opportunity to predict the future trend reversal of stock quotes and to build the rules for it. It gives the possibility to recognize the trend reversal faster than the adaptive methods, which always give a delay of one time step. That's why it helps a trader to make a timely decision about opening and closing positions.