The Application Of Neutral Network And Fuzzy Theory In Forecasting The Stock Price Index

Since the stock market has established more than one hundred years ago, how to predict the stock price accurately has been one of the hottest topics that attracted many scholars' and investors' attention. While there are so many influencing factors that affect the movement of stock prices. All the factors vary in the influence degree, time range and affective mode which results in the exceptional complexity of stock price. So it is very difficult to make some exact forecast of the stock price. Still many scholar and investors insist on the study and practice to find the arbitrage change to achieve the excess proceeds which comes from the accurate forecasting of stock prices. And there are some forecasting methods developed from different position, diverse theory, distinct invest strategy and practice experience.This paper mainly puts up with the application of artificial intelligence and fuzzy theory in the Stock market. Firstly, we present a method for stock price index forecasting using neural network, which involves the BP model and BP-GARCH model, and gives a further discussion on the stock price forecasting application. And the Shenzhen Stock Price Index has been applied forecasting. The empirical result shows that comparing to BP model, the BP-GARCH model has more favorable characteristic such as the convergence rate, learning ability, forecasting precision and estimating error. Secondly, An approach to forecast Stock Index by using interpretable fuzzy models is presented in this paper. Fuzzy clustering algorithm is used to identified initial fuzzy model. The number of fuzzy rules is determined by cluster validity measure. Similar fuzzy sets merging method is then applied to remove redundancy of the fuzzy model to improve its interpretability. In order to obtain high accuracy, yet preserving interpretability, a constrained Levenberg-Marquardt method is utilized to optimize fuzzy model. The proposed approach is applied to ShenZhen Stock Index forecasting process, and results show its validity.