摘要
为了提高大坝安全监控模型的预测精度并检验模型的泛化能力,研究大坝安全监控的统计模型、BP神经网络模型及遗传神经网络模型,并提出基于这两种神经网络的融合模型,结合某拱坝长期的变形观测数据,对上述几种模型进行试算。分析结果表明,所建立的融合模型与其他模型相比具有较高的预测精度,且泛化能力较强,具有良好的适用性。
In order to improve the prediction accuracy and test the generalization ability of the dam safety monitoring model,the statistical model,the BP neural network model and the genetic neural network model are studied.The merging models are proposed based on these two kinds of neural network.According to the long-term deformation observation data of an arch dam,the models mentioned above are calculated.The results show that the merging models have a higher prediction accuracy and a better generalization ability compared to the other models.
出处
《测绘工程》
CSCD
2015年第1期53-56,共4页
Engineering of Surveying and Mapping
基金
江苏省普通高校研究生科研创新计划项目(CXLX11_0143)
关键词
大坝安全监控
神经网络
泛化能力
融合模型
遗传算法
dam safety monitoring
neural network
generalization ability
merging model
genetic algorithm
