With economic development and people's awareness of changes in investment, the stock market has become an important part in modern life. Well the prediction theory of the securities also will be an important value showing.Contains large amounts of real-time stock market data that changes affecting stock price changes by many factors, with great ambiguity, difficult to build accurate models. The internal structure of the system complexity, variability of external factors to make the securities market is positioned as a complex nonlinear dynamic system. Especially short-term forecast of the stock market prediction theory has become a problem in the field, while the traditional time series forecasting methods as well as neural network prediction results are not satisfactory.This dissertation research of neural network and fuzzy system based on the advantages and disadvantages. It made the Cloud theory combine of the neural networks, established securities market prediction based on theoretical models of cloud. The simulation analysis results have been good prediction, thus confirming the cloud theory is applied to predict the feasibility and accuracy of stock market. Membership degree in fuzzy theory, the cloud theory, the concepts of entropy and hyper entropy generator formed by the neural network were the original variable definition and analysis of quantitative, to improve the yield of securities investors provided a favorable theoretical basis and practical tools. The main contents are as follows:(1) This understanding of the stock market's main forecasting methods, and traditional time series forecasting method is not the main drawbacks of the fuzzy input variables and random processing. Allows us to control the introduction of smart stock market.(2) In this paper, neural network technology in the Stock Market is described in detail. Since the expression of neural network input data can not correctly fuzziness and randomness, we attempts to use fuzzy neural network to predict. The simulation results are compared with the actual market behavior. The use of the fuzzy neural network prediction of the feasibility of the Securities and Deficiency try to build cloud theory and neural network forecasting method of combining. Use cloud theory to forecast stock, to overcome the disadvantages of the traditional Fuzzy theory "fuzzy incomplete".(3)The cloud theory itself is a development of fuzzy theory, therefore reserves the grasp of ambiguity. At the same time, use the specific characteristics of the cloud theory, the expected value, entropy, ultra-entropy, reflecting the random nature of stock market. The impact of stock market trends and random characteristics of fuzzy feature fully integrated together, in series with the output of neural network, the neural network information processing system for single-step stock market trend forecast. Predicted and the actual behavior of stock market reflects the comparative analysis of prediction methods predict short-term forecasts of the good results, the Cloud theory and neural network forecasting method of combining the feasibility of forecasting in the stock market verified. In practice, this prediction method to predict stock market provides a reference tool. |