Scholarlink

<p>Concrete mix design is a multi-dimensional task that benefits from robust predictive modeling. This study proposes a weighted ensemble framework comprising Linear Regression, SVR, Random Forest, Gradient Boosting, XGBoost, and a Multi-Layer Perceptron. The ensemble achieved an <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41939_2025_920_Article_IEq1.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="21" /> </InlineMediaObject> <EquationSource Format="TEX">\(R^2\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mi>R</mi> <mn>2</mn> </msup> </math></EquationSource> </InlineEquation> of <b>0.998</b> and an RMSE of <b>0.384&#xa0;MPa</b>, significantly outperforming individual models. Moreover, the proposed SHAP-based interpretation confirmed the critical influence of factors such as binder intensity and water-to-binder ratio on final compressive strength. These results indicate that an intelligently weighted ensemble can meaningfully reduce design complexities for concrete mixtures composed of the studied input variables, offering rapid and accurate strength estimates within the defined scope of mix proportions.</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

An enhanced weighted ensemble approach for predicting concrete compressive strength

  • 期刊论文

  • 发表日期：

    2025年5月28日

查看全文
  • Rupesh Kumar Tipu,
  • Vandna Batra,
  • Suman,
  • Kartik S. Pandya

摘要

Concrete mix design is a multi-dimensional task that benefits from robust predictive modeling. This study proposes a weighted ensemble framework comprising Linear Regression, SVR, Random Forest, Gradient Boosting, XGBoost, and a Multi-Layer Perceptron. The ensemble achieved an \(R^2\) R 2 of 0.998 and an RMSE of 0.384 MPa, significantly outperforming individual models. Moreover, the proposed SHAP-based interpretation confirmed the critical influence of factors such as binder intensity and water-to-binder ratio on final compressive strength. These results indicate that an intelligently weighted ensemble can meaningfully reduce design complexities for concrete mixtures composed of the studied input variables, offering rapid and accurate strength estimates within the defined scope of mix proportions.