Scholarlink

<p>This study is on the application of ensemble machine learning techniques in stone column bearing capacity prediction, integrating experimental and field test data across various soil conditions to address the limitations of traditional methods. This paper developed and evaluated several predictive models, including linear regression, gradient boosting, support vector regression, K-nearest neighbors, and a newly proposed ensemble model. The results revealed that traditional linear regression produced moderate accuracy (R<sup>2</sup> = 0.848, RMSE = 19,220.12, MAE = 14,449.63), while support vector regression substantially underperformed (R<sup>2</sup> = 0.169, RMSE = 53,378.71, MAE = 20497.13), underscoring the challenges posed by data heterogeneity and nonlinearity. In contrast, ensemble approaches, exemplified by gradient boosting (R<sup>2</sup> = 0.986, RMSE = 5806.08, MAE = 1697.49) and K-nearest neighbors (R<sup>2</sup> = 0.987, RMSE = 5593.26, MAE = 1668.71), demonstrated remarkable improvements in accuracy. K-nearest neighbors emerged as the clear front-runner among the models tested, achieving the highest explained variance alongside the lowest error metrics. Its predictions were exceptionally stable; over 90% of residuals fell within ± 10% of actual capacities across all load ranges, and both training and validation R<sup>2</sup> converged above 0.98 as the dataset expanded, indicating minimal overfitting. This combination of peak accuracy, consistency under diverse soil conditions, and strong generalization made K-nearest neighbors the preferred choice for rapid, reliable estimation of stone-column bearing capacity in practical geotechnical design.</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

Ensemble Machine Learning Models for Estimating Stone Column Load-Bearing Capacity Using Combined Experimental and Field Test Data

  • 期刊论文

  • 发表日期：

    2025年9月24日

查看全文
  • Masoud Nasiri,
  • Ehsan Amiri,
  • Ebrahim Fadaei

摘要

This study is on the application of ensemble machine learning techniques in stone column bearing capacity prediction, integrating experimental and field test data across various soil conditions to address the limitations of traditional methods. This paper developed and evaluated several predictive models, including linear regression, gradient boosting, support vector regression, K-nearest neighbors, and a newly proposed ensemble model. The results revealed that traditional linear regression produced moderate accuracy (R2 = 0.848, RMSE = 19,220.12, MAE = 14,449.63), while support vector regression substantially underperformed (R2 = 0.169, RMSE = 53,378.71, MAE = 20497.13), underscoring the challenges posed by data heterogeneity and nonlinearity. In contrast, ensemble approaches, exemplified by gradient boosting (R2 = 0.986, RMSE = 5806.08, MAE = 1697.49) and K-nearest neighbors (R2 = 0.987, RMSE = 5593.26, MAE = 1668.71), demonstrated remarkable improvements in accuracy. K-nearest neighbors emerged as the clear front-runner among the models tested, achieving the highest explained variance alongside the lowest error metrics. Its predictions were exceptionally stable; over 90% of residuals fell within ± 10% of actual capacities across all load ranges, and both training and validation R2 converged above 0.98 as the dataset expanded, indicating minimal overfitting. This combination of peak accuracy, consistency under diverse soil conditions, and strong generalization made K-nearest neighbors the preferred choice for rapid, reliable estimation of stone-column bearing capacity in practical geotechnical design.