Scholarlink

<p>This study presents a novel hybrid machine learning (ML) model, gorilla troops optimization with extreme gradient boosting (GTO-XGBoost), for predicting the unconfined compressive strength (UCS) of solid waste-cement–stabilized cohesive soils. By integrating the global search capabilities of the GTO algorithm with the robust predictive power of XGBoost, this current approach represents a significant advancement over conventional empirical methods and standalone ML models. A comprehensive dataset of 474 encompassing soil properties (liquid limit, plastic limit, plasticity index), material proportions (cement and solid waste content), and chemical composition (silicon dioxide, aluminum oxide, calcium oxide) was used for model training. Performance was evaluated using key metrics, including the coefficient of determination (<i>R</i><sup>2</sup>), mean absolute error (MAE), mean squared error (MSE), and root mean squared error (RMSE). Comparative analysis confirmed the superiority of GTO-XGBoost, which achieved <i>R</i><sup>2</sup> values of 0.996 for both training and testing, along with the lowest error metrics (MAE: 0.059 training, 0.072 testing; MSE: 0.009 training, 0.023 testing; RMSE: 0.097 training, 0.151 testing) compared to random forest and standalone XGBoost. Additionally, to enhance model transparency and interpretability, explainable machine learning methods such as SHapley Additive exPlanations, partial dependence plots, and individual conditional expectation plots were employed, revealing that curing age and cement content are the most critical factors influencing UCS.</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

A Novel Hybrid Prediction Model Leveraging Gorilla Troops Optimization for Unconfined Compressive Strength of Solid Waste-Cement–Stabilized Cohesive Soil

  • 期刊论文

  • 发表日期：

    2025年6月21日

查看全文
  • Majid Khan,
  • Abdul Aziz,
  • Laiba Gulaly

摘要

This study presents a novel hybrid machine learning (ML) model, gorilla troops optimization with extreme gradient boosting (GTO-XGBoost), for predicting the unconfined compressive strength (UCS) of solid waste-cement–stabilized cohesive soils. By integrating the global search capabilities of the GTO algorithm with the robust predictive power of XGBoost, this current approach represents a significant advancement over conventional empirical methods and standalone ML models. A comprehensive dataset of 474 encompassing soil properties (liquid limit, plastic limit, plasticity index), material proportions (cement and solid waste content), and chemical composition (silicon dioxide, aluminum oxide, calcium oxide) was used for model training. Performance was evaluated using key metrics, including the coefficient of determination (R2), mean absolute error (MAE), mean squared error (MSE), and root mean squared error (RMSE). Comparative analysis confirmed the superiority of GTO-XGBoost, which achieved R2 values of 0.996 for both training and testing, along with the lowest error metrics (MAE: 0.059 training, 0.072 testing; MSE: 0.009 training, 0.023 testing; RMSE: 0.097 training, 0.151 testing) compared to random forest and standalone XGBoost. Additionally, to enhance model transparency and interpretability, explainable machine learning methods such as SHapley Additive exPlanations, partial dependence plots, and individual conditional expectation plots were employed, revealing that curing age and cement content are the most critical factors influencing UCS.