The Modeling And Prediction Of Multiresolution Wavelet Network And Its Application In Stock Mark

The ups and downs of stock price is a very complex movement and researched as a hotspot from its being found. With the development of the stock investing, its influence become more and more great, and it is acutely required to deeply understand its movement law.First of all, in the text the essentiality and feasibility are analyzed in the mass reflecting the fact in the home and aboard. Base on it, the theory and method of system modeling is used to strengthen the restriction step by step, the abstract and complicated finance system will be translated into a comparatively clear black-box system which can be simulated and approached, at the same time the brief-time validity will be accepted by nature in this course.The basic theory and method discussed in this paper makes use of wavelet neural network for the prediction in time series of stock indices. Moreover, the data of Shenzhen Exchange between 1998 and 2000.In this paper, the theory and method of neural network and wavelet are studied and understood especially in their merit and defect, a wavelet neural network is advanced which may be more suit to the stock-indices series prediction by reference and use reference all kinds of current theory and method on wavelet, and deriving their merit.In order to validate the correctness of the theory and method, the programmer is made in C++ language, and it is used to predict 20 dates in 2000, and the result is satisfactory, its precision is excel to artificial neural network.