智能算法预测股市自指涉系统的探究

The Exploration of Intelligent Algorithm to Predict the Self-reference System of Stock Market

本文运用Matlab灰色关联度分析算法,计算出个股涨跌幅与沪市大盘涨跌幅的关联系数,然后进行排序,并选定一定数量的关联系数较大的股票作为指标股。股票大盘指数的随机过程可以视为在一定时间区域内的灰色过程。同时具备"无后效性"的特点,使用马尔科夫模型处理随机波动大的问题,弥补了灰色理论中核心和基础的GM(1,1)的不足,从而建立起灰色——马尔科夫预测模型解决股票大盘指数的预测问题。通过关联度分析确定的指标股来预测未来股市大盘的走势。

With the aid of Matlab NN Toolbox, this work designed a three layer Back Propagation NN and optimized the network input. Forecast the Shanghai Stock Index after training and simulation with the method of LM al gorithm and regularization. The result show the network converges easily and has well generalization ability