优化参数的GM(1,1)模型组合改进

COMBINATION IMPROVEMENT METHODS FOR PARAMETERS OPTIMIZED GM(1,1) MODEL

针对提高模型的拟合效果问题,在建模的基础上,提出一种新的模型参数估计方法。结合组合预测的思想,利用BP神经网络预测算法对改进后的灰色模型做进一步的优化修正。该模型充分结合了灰色系统和人工神经网络的各自优势,弥补了单一模型预测精度低的特点,适当扩大了模型的应用范围。

In connection with the improvement of the fitting effect of modeling, this paper puts forward the improvement scheme for pa- rameters based on GM（ 1,1 ） model. In this paper, according to combination prediction method , the improved prediction model based on GM（ 1,1 ） is revised by the neural network, which take advantages of the superiority of GM（ 1,1 ） model and neural network. This method overcome the character of low focasting accuracy of a individual forecasting model.