Computerized profitunity approach based on fuzzy neural networks to shanghai stock market

Stock market moving direction prediction is equivalent to a problem of pattern recognition. Just like the riverbed determines the behavior of a river, Profitunity Approach considers that the market behavior is determined by its structure that can be examined through the market behavior: the changes of price and trading volume. First, the analysis of the conceptual equivalence of the structure characterized by Profitunity Approach and the reconstructed attractor of a nonlinear system is first given. Then, the tools designed by profitunity approach to observe the market structure are transformed to the computable variables by computer. These variables are employed to constitute the feature pattern vectors that well represent the market structure. Finally, The pattern vectors are taken as inputs to the Adaptive-Network-based Fuzzy Inference System to carry out the pattern recognition. The hybrid algorithm of gradient decent and least squares are adopted to tune the fuzzy rules fine and one day future moving direction of Shanghai stock market is forecasted with classifying the rise or fall in future of Shanghai Stock Composite Index.