基于机器学习模型理论的基坑沉降预测研究

Based on machine learning model theory research on the prediction of foundation pit settlement

基坑开挖引起的地面沉降是基坑工程中需要注意的关键性问题,其具有出现区域广泛、发生灾害后果恶劣等特点。为了对基坑沉降进行准确有效的预测,以杭州市某基坑为工程背景,分别采用LSTM(Long Short Term Memory)神经网络和SVM(Support Vector Machines)对基坑沉降建立了预测模型,并通过预测任务评价指标和散点图误差线来检验模型的预测精度。研究结果表明:LSTM模型相比SVM模型表现出了更高的预测精度,更适用于基坑沉降的预测问题,并可为施工现场提供可靠的理论参考。

Ground subsidence caused by foundation pit excavation is a key issue that needs to be paid attention to in foundation pit engineering.It has the characteristics of a wide occurrence area and severe disaster consequences.To accurately and effectively predict the foundation pit settlement,this paper takes a certain foundation pit in Hangzhou as the engineering background,uses LSTM(Long Short Term Memory)neural networ k and SVM(Support Vector Machines)to establish a prediction model for the foundation pit settlement,and uses the prediction task evaluation index and the scatter plot error line To test the prediction accuracy.The research results show that the LSTM model has higher prediction accuracy than the SVM model,is more suitable for the prediction of foundation pit settlement,and can provide a reliable theoretical reference for the construction site.