Fuzzy Neural Network In The Research On The Application Of Stock Forecast

As China’s economic system reform and the deepening of the reform of the financial system has become a stock investment, an important part of social life. Forecasting stock prices as investors focused on care and research. But the stock market is an extremely complex nonlinear dynamic system with high noise, severe nonlinear and investors blindly any factors and other factors as well as the correlation between the complex, causing its price fluctuations tend to show a more strong nonlinearity. In addition, the stock market modeling and forecasting the amount of information is often handled by a very large and very demanding on the algorithm, it is precisely because of its nonlinear characteristics of complex, making prediction about the stock market is often difficult to satisfactory results. How to build a computing speed and accuracy higher than the stock market forecasting model for financial investors with the theoretical and practical-valueIn this paper, the stock price forecasting difficulties of fuzzy logic and neural network concepts, the use of fuzzy logic can be fuzzy natural language knowledge and performance can be Max, Min knowledge of such a simple operation to achieve the characteristics of fuzzy inference, and the use of neural networks to generate knowledge of the rules do not clearly demonstrate its powerful self-learning ability and the characteristics of the two together constitute the fuzzy neural network. TS fuzzy rules and the use of feed-forward neural network approach to modeling, and discusses the network structure, the principles set number of hidden nodes, the sample selection and processing of data, initial parameter determination and other issues.According to the fuzzy neural network stock prediction principle, a fuzzy neural network based stock market forecasting model, using correlation analysis of the stock projections were screened for entry. By MATLAB7.0software, select the Greenview real estate, Weichai Power, China Merchants Securities, Baosteel and empirical analysis on the Shanghai Composite Index, according to the forecast performance of neural network used to predict the results of the evaluation index were evaluated, confirmed the fuzzy neural network prediction is valid, forecasting system is a success.