摘要
针对现有的基坑监测和预测数据处理方法的不足,对BP神经网络预测模型作了研究和改进,应用改进后的BP算法对基坑支护结构水平位移数据进行处理,并将改进的BP算法与传统算法所建立的模型应用于工程实例进行比较,结果表明,改进后的BP神经网络模型在预测精度、训练时间、稳定性等方面均优于传统BP神经网络模型。
Due to the disadvantages of current methods for monitoring and surveying horizontal displacement, this paper puts forward a new meth od, namely improving backpropagation neural networks in measuring the horizontal displacement of foundation pit supports. Through the applica tion in actual projects, compared with the traditional backpropagation neural networks, the improved backpropagation neural networks prove su perior in forecast accuracy, training time, stability and so on.
出处
《山西建筑》
2014年第10期57-58,共2页
Shanxi Architecture
关键词
基坑
水平位移
改进BP神经网络
foundation pit, horizontal displacement, improve backpropagation neural networks
