灰色Verhulst-BP网络组合模型在预测中的应用研究

摘要

Based on the grey Verhulst model and BP-neural network theory,this paper studies the combination of them,puts forward the partial-data Verhulst model group and proposes a kind of gray Verhulst-BP neural network combined forecasting model.In this paper, BP-neural network is utilized to build up the nonlinear mapping between partial-data Verhulst model group and original data, and the defects of the neural networking training with small sample of time series data are over-come.The experiment results show that the combined forecasting model is valid,high precision and good stability.%在灰色Verhulst模型和BP神经网络理论的基础上,对两者的结合方式进行了研究,提出了部分数据Verhulst模型组的概念,得到了一种结合灰色Verhulst与BP神经网络的组合预测模型,利用BP神经网络建立部分数据Verhulst模型组与原始数据之间的非线性映射关系,克服了小样本时间序列数据在神经网络训练时的缺陷.实验结果和仿真验证表明,该组合预测模型具有较高的预测精度和良好的稳定性.