时间序列模型和神经网络模型在股票预测中的分析

利用MATLAB软件编程建立AR模型、RBF和GRNN神经网络模型,滚动预测上证指数开盘价、最高价、最低价和收盘价与实际价格对比,分析误差.结果表明,3种模型用于股票预测均是可行的,误差很小.AR模型不稳定,对个别预测较准;RBF和GRNN网络训练速度都很快,但GRNN比RBF预测效果好.

Analysis of AR Model and Neural Network for Forecasting Stock Prices

Three types of price forecasting models based on AR model、RBF and GRNN neural network are developed by using MATLAB software,to predict the opening price,the highest price,low and closing price of the finaly 5 days by rolling prediction and compare them with the actual price to analyze error.As a result,the three kinds of models execute close prediction,error is small.AR model is instable,the more potential for individual prediction; both RBF and GRNN training speed are fast,but the prediction of GRNN network is better than RBF network.