摘要
将主成分分析与BP神经网络相结合应用到大坝变形影响因子的优化中,建立大坝变形预测模型。可以有效地降低输入因子的维数,减小因子之间相关性的影响,简化网络结构,降低网络训练难度,提高预测的稳定性及精度,提升BP网络训练的效率,解决由影响因子内部相关性而需引入大量因子的问题。通过实验结果对比表明,主成分分析与BP网络相结合的大坝预报模型精度及稳定性明显优于其他模型。
The principal component analysis with the BP neural network is applied to the optimization of dam deformation impact factors, build the dam deformation forecast model. It can effectively reduce the input factor of dimensionality, simplify network structure, reduce network training difficulties, improve the stability and precision, reduce the correlation between the impact of factor and the efficiency of BP network training is improved, a large number of factors are introduced by the correlation between factors are resolved. The result shows that the accuracy and stability are predicted by the application of BP neural network based on principal component analysis is better than other models.
出处
《东华理工大学学报（自然科学版）》
CAS
2011年第3期288-292,共5页
Journal of East China University of Technology(Natural Science)
基金
江西省教育厅重点科技项目(GJJ10022)
关键词
主成分分析
大坝变形
BP神经网络
the principal component analysis
dam deformation
BP neural network
