改进BP神经网络及其在西北建筑业预测中的应用

BP神经网络是分析处理复杂非线性问题的一种有效方法,是目前广泛应用的一种神经网络,已被逐渐应用于对宏观经济问题的研究中。本文有机地整合了计量经济学与BP神经网络,建立了基于因果关系理论来确定BP网络的输入变量,基于协整理论来分析BP网络系统的可靠性,基于学习率可变的动量BP算法的用于研究经济领域问题的改进BP神经网络预测模型,加强了网络模型的理论基础,提高了网络模型的质量,并将其应用于西北建筑业的预测和控制中,取得了令人满意的效果。

Improvement of BP Neural Network and Application in Predicting Northwest Construction Industry

BP neural network is an effective method to deal with complex and nonlinear problems. It has been widely used and gradually applied to deal with problems of macroeconomics at present. In this paper, the author systematically combines econometrics with BP neural network,and improves BP neural network predictive model for research economics field problems, which is based on causality theories to determine the input variable of BP neural network, on coitegration theories to analyze the reliability of BP neural network system, and on the momentum BP algorithm of alterable learn rate. It strengthens the theoretical basis and improves the quality of network model. The improved BP neural network has been applied to predict and control northwest construction industry and the results are satisfactory.