<p>The main goal of this study is to predict the maximum dry density (MDD) and optimal moisture content (OMC) of soil inh building foundation applications using machine learning (ML) techniques, notably Decision Tree Regression (DTR) and Support Vector Regression (SVR) models. The Coati Optimization Algorithm (COA) was incorporated into the study with the goal of improving the precision of MDD and OMC predictions in order to improve the predictive power of these models. By incorporating these ML models into the optimization strategies of COA, new hybrid models were formed that achieved a greater degree of precision within the forecasts. Results in this study illustrate that through validation, DTCO performs well in the OMC target compared to DTR. In the DTCO model, the RMSE shows 0.973, and its <InlineEquation ID="IEq1"> <EquationSource Format="TEX">${R}^{2}$</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mi>R</mi> </mrow> <mn>2</mn> </msup> </math></EquationSource> </InlineEquation> value is 0.983, whereas in the DTR, the error value is 1.726, and the <InlineEquation ID="IEq2"> <EquationSource Format="TEX">${R}^{2}$</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mi>R</mi> </mrow> <mn>2</mn> </msup> </math></EquationSource> </InlineEquation> is 0.942. Similarly, the SVCO model, when pitched against the SVR model, was found to be more accurate in predicting OMC. Since then, the SVCO model outperformed the SVR model with an RMSE of 1.726 and <InlineEquation ID="IEq3"> <EquationSource Format="TEX">${R}^{2}$</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mi>R</mi> </mrow> <mn>2</mn> </msup> </math></EquationSource> </InlineEquation> value of 0.942 for the SVCO against 2.128 and 0.911, respectively, for the SVR model. These results demonstrate the potential of hybrid models to increase prediction accuracy in civil engineering by highlighting how integrating COA optimization approaches with ML models may improve the accuracy and dependability of MDD and OMC forecasts in soil foundation building.</p>

Predicting maximum dry density and optimum moisture content of soil in foundation construction using machine learning algorithms: a comprehensive analysis of input factors

期刊论文
发表日期：

2025年9月16日

查看全文

Yubian Wang,
Rui Li,
Xurong Chu,
Yin Cai

摘要

The main goal of this study is to predict the maximum dry density (MDD) and optimal moisture content (OMC) of soil inh building foundation applications using machine learning (ML) techniques, notably Decision Tree Regression (DTR) and Support Vector Regression (SVR) models. The Coati Optimization Algorithm (COA) was incorporated into the study with the goal of improving the precision of MDD and OMC predictions in order to improve the predictive power of these models. By incorporating these ML models into the optimization strategies of COA, new hybrid models were formed that achieved a greater degree of precision within the forecasts. Results in this study illustrate that through validation, DTCO performs well in the OMC target compared to DTR. In the DTCO model, the RMSE shows 0.973, and its ${R}^{2}$ $R^{2}$ value is 0.983, whereas in the DTR, the error value is 1.726, and the ${R}^{2}$ $R^{2}$ is 0.942. Similarly, the SVCO model, when pitched against the SVR model, was found to be more accurate in predicting OMC. Since then, the SVCO model outperformed the SVR model with an RMSE of 1.726 and ${R}^{2}$ $R^{2}$ value of 0.942 for the SVCO against 2.128 and 0.911, respectively, for the SVR model. These results demonstrate the potential of hybrid models to increase prediction accuracy in civil engineering by highlighting how integrating COA optimization approaches with ML models may improve the accuracy and dependability of MDD and OMC forecasts in soil foundation building.