Some Predict The Price Of The Securities And Empirical Research

As stock market is a kind of complex non-linear dynamic system,the prediction results of traditional prediction technology are unsatisfactory.Neural Network is an estimator of adaptive resonance function,which is not dependent on models,and could fulfill arbitrarily functional relation without them.At present,Neural Networks are becoming a more powerful tool in predicting and establishing non-linear dynamic system. In this dissertation,we mainly study the prediction methods and improve the efficiency from wavelet neural networks,grey model and RBF neural networks.Then the methods are applied to Shanghai stock market.This dissertation consists of the following parts:At first,this paper introduced stock index and the timing of the background research, and the stock timing model in the process of establishing the necessary basic theory and basic theoretical concepts.Secondly,we have verified the existence of non-linear characteristics of the Shanghai securities market.We have in-depth study of multivariate time series and dependent variables measure applied to the timing of stock data。This chapter will also WNN and component analysis,to promote the use of multivariable time series prediction,and to permit the use of transport index data for simulation.Thirdly,based on the chaotic time series in the securities market,this paper studies the principle and deficiency in securities market forecasting by the grey model,RBF neural networks and linear combination forecast,and introduces a new method.Two models(GM(1,1)and RBF neural networks) and wavelet neural net-works were combined to propose a new combination forecasting model for forecasting on Composite Stock Price Index of the stock market in shanghai,China.Finally,some problems in the present research of chaotic prediction and the direction for further study in this field are illuminated.