Scholarlink

<p>Surface settlement during subway construction reflects ground disturbance. Predicting it accurately is challenging because of complex geological and construction conditions. In this study, construction report data were used to create a dataset, and relevant parameters and settlement data were extracted. Three machine learning models (i.e., backpropagation neural network, support vector machine, and extreme gradient boosting tree (XGB)) were used to predict settlement under different construction and geological conditions. The hyperparameters of the models were determined using Bayesian optimization (BO). The contribution of features to the prediction results for the three machine learning models was quantified using the Shapley additive explanation (SHAP), and the performance of the test set under different numbers of features determined by SHAP feature importance was evaluated. Finally, BO-XGB was identified as the optimal machine learning model for predicting surface settlement caused by subway construction.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Study on the Prediction of Subway Construction Settlement Based on Machine Learning

  • 期刊论文

  • 发表日期：

    2025年11月1日

查看全文
  • Xingzhong Nong,
  • Yanmei Ruan,
  • Junsheng Chen,
  • Yuefeng Wu,
  • Yuehua Liang,
  • Wentian Xu

摘要

Surface settlement during subway construction reflects ground disturbance. Predicting it accurately is challenging because of complex geological and construction conditions. In this study, construction report data were used to create a dataset, and relevant parameters and settlement data were extracted. Three machine learning models (i.e., backpropagation neural network, support vector machine, and extreme gradient boosting tree (XGB)) were used to predict settlement under different construction and geological conditions. The hyperparameters of the models were determined using Bayesian optimization (BO). The contribution of features to the prediction results for the three machine learning models was quantified using the Shapley additive explanation (SHAP), and the performance of the test set under different numbers of features determined by SHAP feature importance was evaluated. Finally, BO-XGB was identified as the optimal machine learning model for predicting surface settlement caused by subway construction.