Research On Stock Analysis And Prediction Method Of Nonlinear Approaching Bases On Wavelet Network

In the world, analyzed and predict stock is abroad to use in the investment boundas the rapid development of stock market. In the recent years, using nonlinearityconfirm system rule to research share price action is display more and more life-forcebecause the Chaos and Fractal theory is intermittent produced.In the text, firstly the author introduce the development general situation of Chinastock market, secondly the author expatiate on the fundamental, basic framework andbasic function of BP Neural Network and RBF Neural Network, thirdly the authorintroduce the principle of the Wavelet Analysis and it combine with the RBF NeuralNetwork theory, and discuss the Wavelet Network apply to analyzed and predict Shareprice probability and usability base on theory, finally it judge stock marketdevelopment trend even make out short-team analysis and prediction through trainingthe conformation Wavelet Network. The bring forward method is great assist to solveprediction error rate quite high as well as predict veracity low, arithmeticconstringency slow and soon on which are produced by existence method.To prove this algorithmic validity, the paper random select Enlist Business Bankissue H stock on November 7, 2005 and 21 branch single stock data of Shanghaiwhich come into the market in different time sect in 2007 to put up network trainingand forecast test. The results of test experiments show that Wavelet Networknonlinearity approach analysis and prediction method can predict the single shareprice about 80 percent average exactness rate and it also excelled traditional modeland analysis arithmetic.