Research On The Stock Price Prediction

With the development of Chinese economy, the stock market has been gradually improved. And more and more people put their momey into the stock market and long to know the stock price in advance at a high degree of accuracy as possible. This issue has been focused ever since the stock market comes into being and many scholors have tried to work out it.The theory and methords on the stock price forecast are presented comprehensively in this thesis in terms of the fundamentals, the basic concepts, brief theory, and the simple usage. Methods include ARIMA, neural networks, gray theory and Support Vector Mechine (SVM), placing emphasis on the last one. For better understanding of these methods, a detailed analysis and comparison of advantages and disadvantages of the methods are made.In addition, an empirical analysis of stock prices based on the Support Vector Mechine theory and ARIMA is carried out through regression models and, thus, prediction of stock price in the last month of 2006 is made after optimal parameters of SVM-based model are found. The result from the two methods shows that SVM-based method is better than ARIMA in forecast of the stocks price.