A Novel MBAS-RF Approach to Predict Mechanical Properties of Geopolymer-Based Compositions DOI Creative Commons
Shuzhao Chen, Mengmeng Zhou, Xuyang Shi

et al.

Gels, Journal Year: 2023, Volume and Issue: 9(6), P. 434 - 434

Published: May 24, 2023

Using gels to replace a certain amount of cement in concrete is conducive the green industry, while testing compressive strength (CS) geopolymer requires substantial effort and expense. To solve above issue, hybrid machine learning model modified beetle antennae search (MBAS) algorithm random forest (RF) was developed this study CS concrete, which MBAS employed adjust hyperparameters RF model. The performance verified by relationship between 10-fold cross-validation (10-fold CV) root mean square error (RMSE) value, prediction evaluating correlation coefficient (R) RMSE values comparing with other models. results show that can effectively tune model; had high R (training set = 0.9162 test 0.9071) low 7.111 7.4345) at same time, indicated accuracy high; NaOH molarity confirmed as most important parameter regarding importance score 3.7848, grade 4/10 mm least parameter, 0.5667.

Language: Английский

Application of optimization‐based regression analysis for evaluation of frost durability of recycled aggregate concrete DOI
Mahzad Esmaeili‐Falak, Reza Sarkhani Benemaran

Structural Concrete, Journal Year: 2024, Volume and Issue: 25(1), P. 716 - 737

Published: Jan. 7, 2024

Abstract Concrete constructed using recycled aggregates in place of natural is an efficient approach to increase the construction sector's sustainability. To improve aggregate concrete () technologies permafrost, it essential certify stability frost‐induced conditions. The main goal this study was use support vector regression methods forecast frost durability on basis agent value cold climates. Herein, three optimization called Ant lion (), Grey wolf and Henry Gas Solubility Optimization were employed for indicating optimal values key parameters. results depicted that all developed models have capability predicting regions. as a comprehensive index model has higher at 0.0571 weakest model, then 0.0312 recognized second‐highest finally system 0.0224 mentioned outperformed model. approaches likewise capable accurately forecasting regions, but created method them when taking into account explanations justifications.

Language: Английский

Citations

48
Predicting Penetration Depth in Ultra-High-Performance Concrete Targets under Ballistic Impact: An Interpretable Machine Learning Approach Augmented by Deep Generative Adversarial Network DOI Creative Commons
Majid Khan,

Muhammad Faisal Javed,

Nashwan Adnan Othman

et al.

Results in Engineering, Journal Year: 2025, Volume and Issue: unknown, P. 103909 - 103909

Published: Jan. 1, 2025

Language: Английский

Citations

2
A Comprehensive Survey on Arithmetic Optimization Algorithm DOI Open Access
Krishna Gopal Dhal, Buddhadev Sasmal, Arunita Das

et al.

Archives of Computational Methods in Engineering, Journal Year: 2023, Volume and Issue: 30(5), P. 3379 - 3404

Published: March 15, 2023

Language: Английский

Citations

31
Estimation of unconfined compressive strength of marine clay modified with recycled tiles using hybridized extreme gradient boosting method DOI
Daihong Li, Xiaoyu Zhang, Qian Kang

et al.

Construction and Building Materials, Journal Year: 2023, Volume and Issue: 393, P. 131992 - 131992

Published: June 10, 2023

Language: Английский

Citations

30
Data-driven estimates of the strength and failure modes of CFRP-steel bonded joints by implementing the CTGAN method DOI
Songbo Wang, Tim Stratford, Yang Li

et al.

Engineering Fracture Mechanics, Journal Year: 2024, Volume and Issue: 299, P. 109962 - 109962

Published: Feb. 20, 2024

Language: Английский

Citations

14
Soil–Structure Interaction for Buried Conduits Influenced by the Coupled Effect of the Protective Layer and Trench Installation DOI
Ebrahim Hassankhani, Mahzad Esmaeili‐Falak

Journal of Pipeline Systems Engineering and Practice, Journal Year: 2024, Volume and Issue: 15(2)

Published: Feb. 24, 2024

Language: Английский

Citations

12
Uncertainty quantification in multiaxial fatigue life prediction using Bayesian neural networks DOI
GaoYuan He, Yongxiang Zhao,

ChuLiang Yan

et al.

Engineering Fracture Mechanics, Journal Year: 2024, Volume and Issue: 298, P. 109961 - 109961

Published: Feb. 15, 2024

Language: Английский

Citations

11
Artificial intelligence techniques in advanced concrete technology: A comprehensive survey on 10 years research trend DOI Creative Commons
Ramin Kazemi

Engineering Reports, Journal Year: 2023, Volume and Issue: 5(9)

Published: May 23, 2023

Abstract Advanced concrete technology is the science of efficient, cost‐effective, and safe design in civil engineering projects. Engineers designers are generally faced with slightest change conditions or objectives project, which makes it challenging to choose optimal among several ones. Besides, experimental examination all them requires time high costs. Hence, an efficient approach utilize artificial intelligence (AI) techniques predict optimize real‐world problems technology. Despite large body publications this field, there few comprehensive surveys that conduct scientometric analysis. This paper provides a state‐of‐the‐art review lists, summarizes, categorizes most widely used machine learning methods, meta‐heuristic algorithms, hybrid approaches issues. To end, 457 considered during recent decade highlight annual trend/active journals/top researchers/co‐occurrence key title words/countries' participation/research hotspots. In addition, AI classified into distinct clusters using VOSviewer clustering visualization identify application scope their relationship through link strength. The findings can be beacon help researchers future research on advanced

Language: Английский

Citations

21
Hybrid BO-XGBoost and BO-RF Models for the Strength Prediction of Self-Compacting Mortars with Parametric Analysis DOI Creative Commons
Asif Ahmed, Wei Song, Yumeng Zhang

et al.

Materials, Journal Year: 2023, Volume and Issue: 16(12), P. 4366 - 4366

Published: June 13, 2023

Self-compacting mortar (SCM) has superior workability and long-term durable performance compared to traditional mortar. The strength of SCM, including both its compressive flexural strengths, is a crucial property that determined by appropriate curing conditions mix design parameters. In the context materials science, predicting SCM challenging because multiple influencing factors. This study employed machine learning techniques establish prediction models. Based on ten different input parameters, specimens were predicted using two types hybrid (HML) models, namely Extreme Gradient Boosting (XGBoost) Random Forest (RF) algorithm. HML models trained tested experimental data from 320 test specimens. addition, Bayesian optimization method was utilized fine tune hyperparameters algorithms, cross-validation partition database into folds for more thorough exploration hyperparameter space while providing accurate assessment model's predictive power. results show can successfully predict values with high accuracy, Bo-XGB model demonstrated higher accuracy (R2 = 0.96 training R2 0.91 testing phases) low error. terms prediction, BO-RF performed very well, train 0.88 stages minor errors. Moreover, SHAP algorithm, permutation importance leave-one-out score used sensitivity analysis explain process interpret governing variable parameters proposed Finally, outcomes this might be applied guide future

Language: Английский

Citations

21
A Review of the Data-Driven Prediction Method of Vehicle Fuel Consumption DOI Creative Commons

Dengfeng Zhao,

Haiyang Li, Junjian Hou

et al.

Energies, Journal Year: 2023, Volume and Issue: 16(14), P. 5258 - 5258

Published: July 9, 2023

Accurately and efficiently predicting the fuel consumption of vehicles is key to improving their economy. This paper provides a comprehensive review data-driven prediction models. Firstly, by classifying summarizing relevant data that affect consumption, it was pointed out commonly used currently involve three aspects: vehicle performance, driving behavior, environment. Then, from model structure, predictive energy characteristics traditional machine learning (support vector machine, random forest), neural network (artificial deep network), this point that: (1) based on networks has higher processing ability, training speed, stable ability; (2) combining advantages different models build hybrid for prediction, accuracy can be greatly improved; (3) when comparing indicts, both method consistently exhibit coefficient determination above 0.90 root mean square error below 0.40. Finally, summary prospect analysis are given various models’ performance application status.

Language: Английский

Citations

18