Scholarlink

<p>This paper presents Multiple Linear and Machine Learning models of bearing capacity for strip footings at sandy clay slopes subjected to vertical loads. Several parameters are considered in the analysis, including footing width, embedment depth, unit weight, slope angle, internal friction angle, and soil cohesion. A finite element analysis is conducted to assess the impact of these factors. Additionally, an empirical prediction for bearing capacity is proposed. Machine learning techniques utilising various models are employed to analyse performance outcomes, with the Shapley Additive Explanations (SHAP) method used to quantify the contribution of each parameter. The results show that the empirical formulation for predicting ultimate bearing capacity can be effectively applied in engineering practice. Significantly, the findings indicate that the XGBoost model yields the most precise predictions of bearing capacity. The primary parameters influencing bearing capacity include embedded depth, width, unit weight, and internal friction angle, whereas vertical load and unit weight have a minimal impact.</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

Integrating Multiple Linear Regression Analysis and Machine Learning Models to Predict the Bearing Capacity of Strip Footings on Sandy Clay Slopes

  • 期刊论文

  • 发表日期：

    2025年2月5日

查看全文
  • Lindung Zalbuin Mase,
  • Rena Misliniyati,
  • Nia Afriantialina Muharama,
  • Fepy Supriani,
  • Debby Ariansyah Ahmad,
  • Ryan Fernanda,
  • Vinay Bhushan Chauhan,
  • Salisa Chaiyaput

摘要

This paper presents Multiple Linear and Machine Learning models of bearing capacity for strip footings at sandy clay slopes subjected to vertical loads. Several parameters are considered in the analysis, including footing width, embedment depth, unit weight, slope angle, internal friction angle, and soil cohesion. A finite element analysis is conducted to assess the impact of these factors. Additionally, an empirical prediction for bearing capacity is proposed. Machine learning techniques utilising various models are employed to analyse performance outcomes, with the Shapley Additive Explanations (SHAP) method used to quantify the contribution of each parameter. The results show that the empirical formulation for predicting ultimate bearing capacity can be effectively applied in engineering practice. Significantly, the findings indicate that the XGBoost model yields the most precise predictions of bearing capacity. The primary parameters influencing bearing capacity include embedded depth, width, unit weight, and internal friction angle, whereas vertical load and unit weight have a minimal impact.