遗传程序设计分析股价移动平均及中长期走势*

将遗传程序设计应用到股票价格分析，在股票市场各种因素相互作用与影响很难厘清的情况下，只从个别因素（价格）入手，测试对单一因素预测所能达到的效果；提出了两种预测方法：对不同尺度的股票移动平均线进行预测和对股票价格数据进行平滑预处理之后所进行的中长期预测。通过遗传程序设计算法，寻找前几个时间单位的股票价格对本期股票价格影响的经验公式，以期反映价格变动的规律。计算机实验模拟表明，该方法对于平均线的预测和中长期预测有较好的效果。

Applying genetic programming to analyze moving average and long & mid-term trends of stock prices

This paper employed genetic programming (GP) to analyze stock price. The task tried to find out how far it could go if used only one element, which was the price, to predict the stock market, based on the understanding that it was impossible to distinguish all the interactions between various elements in the stock market. Our work proposed two multi-scale approaches trying to predict stock prices. One was to use GP to form empirical formulas to predict the moving average lines of stock prices; the other was to use GP to do long & mid-term predictions on pre-processed data. The aim was to find empirical laws for specific enterprises stock prices based on previous stock price data. Simulations show that the method to predict the moving average and long & mid-term trends of stock prices is effective.