市政技术2025，Vol.43Issue(5)：222-230,9.DOI:10.19922/j.1009-7767.2025.05.222
海泥砂浆抗压强度的机器学习与可视化研究
Research on Machine Learning and Visualization of the Compressive Strength of Marine Mud Mortar
摘要
Abstract
Under the guidance of the"Maritime Power"and"Dual Carbon Goals",the reuse of marine sediments has become a hot research topic in the industry.Under the background of the rapid development of artificial intelligence(AI),the relevant research has been conducted on machine learning and visualization of the compressive strength of marine mud mortar.Four types of machine learning models,namely Gaussian Process Regression(GPR),Support Vector Regression(SVR),Random Forest(RF),and Extreme Gradient Boosting Decision Tree(XGBoost)are used to train and predict more than 200 sets of collected data in order to adapt to the prediction model of compressive strength of sea mud mortar.SHapley Additive Explanations(SHAP)visualization analysis was performed on the op-timal prediction results to explore the influence of various features on the compressive strength of marine mud mortar.Finally,the optimized XGBoost model prediction results are verified and analyzed through experiments.The experi-mental results show that all the determination coefficients predicted by the training and testing sets of the four types of models are all greater than 0.90,and the root mean square error is less than 4.50,indicating that the training effects are better.Among them,the XGBoost model has the best prediction results,indicating that the XGBoost model is more suitable for predicting the compressive strength of sea mud mortar;The dosage of marine mud has the greatest impact on the compressive strength of marine mud mortar with a negative correlation;The compressive strength of marine mud mortar increases with the increase of age of marine mud mortar,cement or sand dosage,and shows a trend of first increase and then decrease with the increase of water usage,which is consistent with the results of con-ventional experimental analysis;The use of optimized XGBoost model has high reliability for predicting the com-pressive strength of marine mud mortar.关键词
海泥砂浆/抗压强度/机器学习/SHAP可视化Key words
marine mud mortar/compressive strength/machine learning/SHAP visualization分类
建筑与水利引用本文复制引用
周马技,郑佳凯,黄任,梁科仁,彭博远,谢蕙竹,李古..海泥砂浆抗压强度的机器学习与可视化研究[J].市政技术,2025,43(5):222-230,9.基金项目
广东省基础与应用基础研究基金(2024A1515030088) （2024A1515030088）
广州市科技计划项目(2023A03J0088) （2023A03J0088）
大学生创新项目(XJ202311078116) （XJ202311078116）
