摘 要
股票市場是一個風險和利益共存的市場,股票市場的建模和預測研究對我國的經濟發展和金融建設具有重要意義。掌握好股票預測能力，就可以更好地選擇買賣時機，獲得更大的利益。
人工神經網絡具有廣泛的適應能力, 學習能力和映射能力,在多變量非線性系統的建模和控制方面取得了驚人的成就。針對股票市場的不確定性，神經網絡具有比其他算法更有優勢，預測的結果更加精確，更加有效。
SAS Enterprise Miner簡稱EM，是一個集成的數據挖掘系統,它的運行方式是通過在一個工作空間(workspace)中按照一定的順序添加各種可以實現不同功能的節點，然后對不同節點進行相應的設置，最后運行整個工作流程(workflow)，便可以得到相應的結果。
模型建立中，本文通過1990年12月19日到2009年12月31日上證指數日線數據中的開盤價、最高價、最低價、收盤價、成交量以及成交金額延伸出一些專用指標來預測短期股票的漲跌，得出其中的規律，判斷股票買賣時機，從而應用于股票預測。
關鍵字：股價預測 神經網絡 SAS EM
Abstract
Stock market is a market in which risks and benefits of co-existence. It is very important to stock market modeling and prediction on China's economic and financial. Mastering the predictive power of stock, you can choose better trading opportunities to gain more benefits.
There is a wide range of adaptability, learning ability and mapping capabilities in artificial neural network which has made remarkable achievements in multivariable modeling and control of nonlinear systems. Because of the uncertainty of the stock market, neural network is superior to other algorithms and the predictions are more efficient and accurate.
SAS Enterprise Miner referred to as EM, is an integrated data mining system. It runs through a workspace (workspace). In the workspace, a variety of nodes, which can achieve different functions, can be added in accordance with a certain order. By setting different node, than running the entire workflow (workflow), we can obtain the corresponding results.
In this model, the opening price, highest price, lowest price, closing price, trading volume and transaction value of Shanghai Stock Index Date Line Data from December 19, 1990 to December 31, 2009 are as the input. Some extension of the input can predict the stocks, either ups or downs. We can get the regular pattern, determine stock trading opportunity, than use it in stocks prediction.
Keywords: Stock price prediction; Neural Networks; SAS EM
