摘要
现今常见的变形监测数据处理方法有GM(1,1)模型、BP神经网络模型和径向基神经网络模型(RBF),本文分别采用GM(1,1)模型和RBF网络模型对基坑结构的水平位移量进行预测,并且采用基于方差倒数法的组合模型对上述两种单一模型预测值进行组合,以达到改善预测精度的效果。实验结果表明,组合模型的预测精度和可靠性优于单一模型,说明了该模型的可行性。
Nowadays,the common methods of deformation monitoring data processing are GM(1,1)model and Back Propagation Neural Network and netradial basis function neural network.In this paper,GM(1,1)and radial basis function neural network were respectively used into the prediction of horizontal displacement in pits' structure.In order to improve accuracy effect,the combined model based on the variance reciprocal method was applied to the prediction.The results show,both the accuracy and reliability with the combined model are more superior than the single models.And they also show the feasibility of the combined model.
出处
《北京测绘》
2017年第4期46-49,共4页
Beijing Surveying and Mapping
