The Forecast Of Stock Market Chaos’ Effect Based On The Improved RBF Neural Network

With the development of economic globalization and financial liberalization, the stock market which is the core sect of capital market plays a more and more important role. It is inevitably experiencing "a bubble-steady growth-rapid expanding-the final rupture" stage in the stock market’s step by step operation, which has great impact for development of the macroeconomic. Because the phenomenon of stock market turmoil is difficult to quantify, it is hard to obtain the effective model and satisfactory result.Chaos theory is an emerging discipline since the second half of the20th century, it is a special motion form of nonlinear systems. Chaos refers to still have possible inherent randomness in a deterministic nonlinear system, without any under the influence of external random factors. Based on the complexity, openness and systematic analysis, the traditional model is always ignoring the intrinsic link between the various elements. This paper introduces chaos theory to the study and analysis to find the inherent law to improve the efficiency of the use of resource configuration.Firstly, the paper analyzes the features and chaotic phenomena intrinsic link between the stock market, and describes the development of the origin of chaos. It briefly introduces the characteristics of the stock market chaos effect from the three aspects. Combing the basic principles of reconstruction phases, RBF neural network model’s building has the theoretical foundation. It also describes the methods to identification of the effect of chaos from the perspective of qualitative and quantitative, and also introduces three kinds of chaotic prediction methods.Secondly, it evaluates the traditional RBF neural network learning algorithm for training the future. The algorithm has the advantages of small amount of calculation and short training time, but prediction of overall fitting effect is not ideal. To improve the accuracy of forecasting the stock market chaos effect, it describes the necessity and learning principle to expound the traditional RBF neural network training algorithm.According to the correlation dimension method and lyapunov index method to identify the existence of chaotic effect, if the stock time series is the chaos effect, then one can forecast using the improved RBF neural network learning algorithm for training and test.With the improved RBF neural network algorithm, it uses the empirical analysis by training and test of reconstructing the phase space of the empirical analysis, which is for completing the applicability of the model to predict the ideal. The paper combines parts of the original phase point and the prediction of phase points together, in order to predict the short-term effect of chaos analysis. According to the judgment of chaotic characteristic, finally it is concluded that Chinese stock market has the effect of chaos in the short term the future.