基于改进BP神经网络的组合预测模型设计

Design for Combining Forecasting Based on Advanced BP Neural Network

为提高预测精度,解决非线性组合预测中的困难,利用改进BP神经网络对非线性组合预测模型进行了设计.讨论了模型设立的原则和一般程序,比较其与传统的组合预测方法之间的优劣,并给出实例加以验证.结果显示,基于改进BP神经网络的非线性组合预测模型能够准确描述系统中的非线性,提高预测精度.

To improve accuracy of forecast, the paper utilize advanced BP neural network to design for nonlinear combining forecasting model. Based on the principle and general process we put forward, the forecast effects of BP neural network are compared with traditional combining forecasting methods. The conclusion is that the former can improve the accuracy of forecast by mimicing the nonlinear relationship in complicated system.