基于加权均值生成序列的神经网络GM(1,1)模型

Neural Network GM(1,1) Model Based on Weighting Mean Value Generating Sequence

将加权均值生成和神经网络中的BP算法运用于GM(1,1)模型的建模过程,使得参数的优化估计与改善初始条件有机地结合起来,对GM(1,1)模型进行了改进,并将所得模型应用于具体实例中,预测结果表明,改进后的模型精度得到较大提高。

In this paper,weighting mean value generating and BP calculating method of neural network are applied into GM(1,1) modeling,which integrates the optimized estimate of parameters with improving initial terms organically,and makes GM(1,1) model better.Applying the model into concrete example,and forecasting results suggests the improved model has better precision.