A case study of resilient modulus prediction leveraging an explainable metaheuristic-based XGBoost

Transportation Geotechnics, Journal Year: 2024, Volume and Issue: 45, P. 101216 - 101216

Published: Feb. 18, 2024

Language: Английский

---

Dimensionless Machine Learning: Dimensional Analysis to Improve LSSVM and ANN models and predict bearing capacity of circular foundations

Artificial Intelligence Review, Journal Year: 2025, Volume and Issue: 58(4)

Published: Jan. 30, 2025

Language: Английский

---

Decision tree models for the estimation of geo-polymer concrete compressive strength

Mathematical Biosciences & Engineering, Journal Year: 2023, Volume and Issue: 21(1), P. 1413 - 1444

Published: Jan. 1, 2023

<abstract> <p>The green concretes industry benefits from utilizing gel to replace parts of the cement in concretes. However, measuring compressive strength geo-polymer (CSGPoC) needs a significant amount work and expenditure. Therefore, best idea is predicting CSGPoC with high level accuracy. To do this, base learner super machine learning models were proposed this study anticipate CSGPoC. The decision tree (DT) applied as learner, random forest extreme gradient boosting (XGBoost) techniques are used system. In regard, database was provided involving 259 data samples, which four-fifths considered for training model one-fifth selected testing models. values fly ash, ground-granulated blast-furnace slag (GGBS), Na2SiO3, NaOH, fine aggregate, gravel 4/10 mm, 10/20 water/solids ratio, NaOH molarity input estimate evaluate reliability performance (DT), XGBoost, (RF) models, 12 evaluation metrics determined. Based on obtained results, highest degree accuracy achieved by XGBoost mean absolute error (MAE) 2.073, percentage (MAPE) 5.547, Nash–Sutcliffe (NS) 0.981, correlation coefficient (R) 0.991, R² 0.982, root square (RMSE) 2.458, Willmott's index (WI) 0.795, weighted (WMAPE) 0.046, Bias (SI) 0.054, p 0.027, relative (MRE) -0.014, a²⁰ 0.983 MAE 2.06, MAPE 6.553, NS 0.985, R 0.993, 0.986, RMSE 2.307, WI 0.818, WMAPE 0.05, SI 0.056, 0.028, MRE -0.015, 0.949 model. By importing set into trained 0.8969, 0.9857, 0.9424 DT, RF, respectively, show superiority estimation. conclusion, capable more accurately than DT RF models.</p> </abstract>

Language: Английский

---

Modelling Soil Compaction Parameters Using an Enhanced Hybrid Intelligence Paradigm of ANFIS and Improved Grey Wolf Optimiser

Mathematics, Journal Year: 2023, Volume and Issue: 11(14), P. 3064 - 3064

Published: July 11, 2023

The criteria for measuring soil compaction parameters, such as optimum moisture content and maximum dry density, play an important role in construction projects. On sites, base/sub-base soils are compacted at the optimal to achieve desirable level of compaction, generally between 95% 98% density. present technique determining parameters laboratory is a time-consuming task. This study proposes improved hybrid intelligence paradigm alternative tool method estimating density soils. For this purpose, advanced version grey wolf optimiser (GWO) called GWO (IGWO) was integrated with adaptive neuro-fuzzy inference system (ANFIS), which resulted high-performance model named ANFIS-IGWO. Overall, results indicate that proposed ANFIS-IGWO achieved most precise prediction (degree correlation = 0.9203 root mean square error 0.0635) 0.9050 0.0709) outcomes suggested noticeably superior those attained by other ANFIS models, built standard GWO, Moth-flame optimisation, slime mould algorithm, marine predators algorithm. geotechnical engineers can benefit from newly developed during design stage civil engineering MATLAB models also included parameters.

Language: Английский

---

Advanced Tree-Based Techniques for Predicting Unconfined Compressive Strength of Rock Material Employing Non-Destructive and Petrographic Tests

Materials, Journal Year: 2023, Volume and Issue: 16(10), P. 3731 - 3731

Published: May 15, 2023

The accurate estimation of rock strength is an essential task in almost all rock-based projects, such as tunnelling and excavation. Numerous efforts to create indirect techniques for calculating unconfined compressive (UCS) have been attempted. This often due the complexity collecting completing abovementioned lab tests. study applied two advanced machine learning techniques, including extreme gradient boosting trees random forest, predicting UCS based on non-destructive tests petrographic studies. Before applying these models, a feature selection was conducted using Pearson's Chi-Square test. technique selected following inputs development tree (XGBT) forest (RF) models: dry density ultrasonic velocity tests, mica, quartz, plagioclase results. In addition XGBT RF some empirical equations single decision (DTs) were developed predict values. results this showed that model outperforms prediction terms both system accuracy error. linear correlation 0.994, its mean absolute error 0.113. addition, outperformed DTs equations. models also KNN (R = 0.708), ANN 0.625), SVM 0.816) models. findings imply can be employed efficiently

Language: Английский

---

Machine learning approach for predicting compressive strength in foam concrete under varying mix designs and curing periods

Smart Construction and Sustainable Cities, Journal Year: 2023, Volume and Issue: 1(1)

Published: Nov. 10, 2023

Abstract Efforts to reduce the weight of buildings and structures, counteract seismic threat human life, cut down on construction expenses are widespread. A strategy employed address these challenges involves adoption foam concrete. Unlike traditional concrete, concrete maintains standard composition but excludes coarse aggregates, substituting them with a agent. This alteration serves dual purpose: diminishing concrete’s overall weight, thereby achieving lower density than regular creating voids within material due agent, resulting in excellent thermal conductivity. article delves into presentation statistical models utilizing three different methods—linear (LR), non-linear (NLR), artificial neural network (ANN)—to predict compressive strength These formulated based dataset 97 sets experimental data sourced from prior research endeavors. comparative evaluation outcomes is subsequently conducted, leveraging benchmarks like coefficient determination ( R 2 ), root mean square error (RMSE), absolute (MAE), aim identifying most proficient model. The results underscore remarkable effectiveness ANN evident model’s value, which surpasses that LR model by 36% 22%. Furthermore, demonstrates significantly MAE RMSE values compared both NLR models.

Language: Английский

---

A novel approach to estimate rock deformation under uniaxial compression using a machine learning technique

Bulletin of Engineering Geology and the Environment, Journal Year: 2024, Volume and Issue: 83(7)

Published: June 14, 2024

Language: Английский

---

Predicting the International Roughness Index of JPCP and CRCP Rigid Pavement: A Random Forest (RF) Model Hybridized with Modified Beetle Antennae Search (MBAS) for Higher Accuracy

Computer Modeling in Engineering & Sciences, Journal Year: 2024, Volume and Issue: 139(2), P. 1557 - 1582

Published: Jan. 1, 2024

To improve the prediction accuracy of International Roughness Index (IRI) Jointed Plain Concrete Pavements (JPCP) and Continuously Reinforced (CRCP), a machine learning approach is developed in this study for modelling, combining an improved Beetle Antennae Search (MBAS) algorithm Random Forest (RF) model.The 10-fold cross-validation was applied to verify reliability model proposed study.The importance scores all input variables on IRI JPCP CRCP were analysed as well.The results by comparative analysis showed newly MBAS RF hybrid (RF-MBAS) higher, indicated RMSE R values 0.2732 0.9476 well 0.1863 0.9182 CRCP.The obtained result far exceeds that used traditional Mechanistic-Empirical Pavement Design Guide (MEPDG), indicating great potential proportional corresponding study, including total joint faulting cumulated per KM (TFAULT), percent subgrade material passing 0.075-mm Sieve (P 200 ) pavement surface area with flexible rigid patching (all Severities) (PATCH) which scored higher.

Language: Английский

---

Development of machine learning models for forecasting the strength of resilient modulus of subgrade soil: genetic and artificial neural network approaches

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Aug. 6, 2024

Accurately predicting the Modulus of Resilience (MR) subgrade soils, which exhibit non-linear stress–strain behaviors, is crucial for effective soil assessment. Traditional laboratory techniques determining MR are often costly and time-consuming. This study explores efficacy Genetic Programming (GEP), Multi-Expression (MEP), Artificial Neural Networks (ANN) in forecasting using 2813 data records while considering six key parameters. Several Statistical assessments were utilized to evaluate model accuracy. The results indicate that GEP consistently outperforms MEP ANN models, demonstrating lowest error metrics highest correlation indices (R2). During training, achieved an R2 value 0.996, surpassing (R2 = 0.97) 0.95) models. Sensitivity SHAP (SHapley Additive exPlanations) analysis also performed gain insights into input parameter significance. revealed confining stress (21.6%) dry density (26.89%) most influential parameters MR. corroborated these findings, highlighting critical impact on predictions. underscores reliability as a robust tool precise prediction applications, providing valuable performance significance across various machine-learning (ML) approaches.

Language: Английский

---

CBR of stabilized and reinforced residual soils using experimental, numerical, and machine-learning approaches

Transportation Geotechnics, Journal Year: 2023, Volume and Issue: 42, P. 101080 - 101080

Published: Aug. 10, 2023

Language: Английский

---