Scholarlink

<p>The elastic modulus (<i>E</i>) of rocks is an essential parameter in mining and rock engineering projects, as it directly affects their stability and structural integrity. This study investigates the application of metaheuristic optimization algorithms, specifically Cuckoo Search (CS) and Harris Hawks Optimization (HHO), to fine-tune the hyperparameters of ensemble regression models, including extreme gradient boosting (XGBoost), decision tree (DT), and adaptive boosting (AdaBoost). A dataset of 122 rock samples, including input parameters such as wet density, moisture, dry density, Brazilian tensile strength, and uniaxial compressive strength, was used to predict <i>E</i>. A dataset was split into training and testing datasets with a 70:30 ratio. Model performance was evaluated using metrics like coefficient of determination (<i>R</i><sup>2</sup>), root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage&#xa0;error (MAPE) in %, and severity index (SI). The results show that CS-HHO-optimized models significantly outperformed the unoptimized models, with the optimized stacking model providing superior prediction accuracy for predicting <i>E.</i> Both the unoptimized and optimized Stacking Models exhibited superior performance on the test data. The unoptimized Stacking Model achieved an <i>R</i><sup>2</sup> of 0.980, RMSE of 0.0483, MAE of 0.0223, MAPE of 12.412%, and SI of 0.1737, while the CS-HHO-optimized Stacking Model yielded a similar <i>R</i><sup>2</sup> of 0.980, with slight variations in RMSE (0.0497), MAE (0.0234), MAPE (13.5736%), and SI (0.1786). This study provides a robust predictive framework for rock behavior analysis, contributing to the field of mining and rock engineering project 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

Predicting Elastic Modulus of Rocks Using Metaheuristic-Optimized Ensemble Regression Models

  • 期刊论文

  • 发表日期：

    2025年3月18日

查看全文
  • Niaz Muhammad Shahani,
  • Xigui Zheng,
  • Xin Wei,
  • Yue Wei

摘要

The elastic modulus (E) of rocks is an essential parameter in mining and rock engineering projects, as it directly affects their stability and structural integrity. This study investigates the application of metaheuristic optimization algorithms, specifically Cuckoo Search (CS) and Harris Hawks Optimization (HHO), to fine-tune the hyperparameters of ensemble regression models, including extreme gradient boosting (XGBoost), decision tree (DT), and adaptive boosting (AdaBoost). A dataset of 122 rock samples, including input parameters such as wet density, moisture, dry density, Brazilian tensile strength, and uniaxial compressive strength, was used to predict E. A dataset was split into training and testing datasets with a 70:30 ratio. Model performance was evaluated using metrics like coefficient of determination (R2), root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE) in %, and severity index (SI). The results show that CS-HHO-optimized models significantly outperformed the unoptimized models, with the optimized stacking model providing superior prediction accuracy for predicting E. Both the unoptimized and optimized Stacking Models exhibited superior performance on the test data. The unoptimized Stacking Model achieved an R2 of 0.980, RMSE of 0.0483, MAE of 0.0223, MAPE of 12.412%, and SI of 0.1737, while the CS-HHO-optimized Stacking Model yielded a similar R2 of 0.980, with slight variations in RMSE (0.0497), MAE (0.0234), MAPE (13.5736%), and SI (0.1786). This study provides a robust predictive framework for rock behavior analysis, contributing to the field of mining and rock engineering project design.