机构地区: 惠州学院数学系
出 处: 《数学的实践与认识》 2013年第22期28-37,共10页
摘 要: 针对股价指数特有的波动性,提出了基于灰色残差模型和BP神经网络的股指动态预测方法,并运用多元线性回归模型对两种动态预测结果进行拟合.同时,随机抽取部分上证指数和道琼斯指数的实证研究表明:动态预测模型能及时调整新数据对后续预测的影响,获得了较高的预测精度. According to the peculia volatility of the stock index, we put forward dynamic prediction methods based on the grey residual model and the BP neural network. Appling multiple linear regression model, we combine the two dynamic prediction results. Simultane- ously, we experiment on the random part of Dow Jones industries average index and Shanghai Stock Exchange The experimental results demonstrate that the dynamical prediction model can adjust the subsequent influence of the new data and acquire higher precision of prediction.