The Stock market is an extremely complex non-linear dynamics system, participating in stock investment, income comes with risk coexisting. In order to make a profit, the investors have been exploring its inherent laws, seeks for the effective analysis method and the tool. But the stock price system internal structure's complexity, external factor's polytropy had decided the stock market forecasts which is difficult. The traditional forecast tool has not been able to meet this need. Therefore, the stock market inherent laws' research and the forecast have the extremely important theory significance and the application value.This paper on people's conduct market analysis, from the investor perspective, grasp the pulse of the market. And the existing system of the various indicators, to a large extent rely on investor psychology and enthusiasm through the price-volume relationship reflected in the K-map. Rules will be hidden in the law of K-Line map data. Neural networks are highly nonlinear approximation capability and self-learning, adaptive characteristics, and can be haphazard data to identify the inherent law of the development of the stock market, and its storage in the network specific weights, threshold, with to forecast future trends. The experiment proved that uses the BP neural network modeling, with SOM from the organization network to the data class, both unifies, may makes the good forecast progress to the stock price and the trend. |