Short-term rockburst prediction in underground project: insights from an explainable and interpretable ensemble learning model

Acta Geotechnica, Год журнала: 2023, Номер 18(12), С. 6655 - 6685

Опубликована: Июль 10, 2023

Язык: Английский

---

Short-Term Rockburst Damage Assessment in Burst-Prone Mines: An Explainable XGBOOST Hybrid Model with SCSO Algorithm

Rock Mechanics and Rock Engineering, Год журнала: 2023, Номер 56(12), С. 8745 - 8770

Опубликована: Сен. 2, 2023

Язык: Английский

---

Sustainable Digital Marketing Under Big Data: An AI Random Forest Model Approach

IEEE Transactions on Engineering Management, Год журнала: 2024, Номер 71, С. 3566 - 3579

Опубликована: Янв. 1, 2024

Digital marketing refers to the process of promoting, selling, and delivering products or services through online platforms channels using internet electronic devices in a digital environment. Its aim is attract engage target audiences various strategies methods, driving brand promotion sales growth. The primary objective this scholarly study seamlessly integrate advanced big data analytics artificial intelligence (AI) technology into realm marketing, thereby fostering progression optimization sustainable practices. First, characteristics applications involving vast, diverse, complex datasets are analyzed. Understanding their attributes scope application essential. Subsequently, comprehensive investigation AI-driven learning mechanisms conducted, culminating development an AI random forest model (RFM) tailored for marketing. Subsequent this, leveraging real-world case enterprise X, fundamental customer collected subjected meticulous analysis. RFM model, ingeniously crafted study, then deployed prognosticate anticipated count prospective customers said enterprise. empirical findings spotlight pronounced prevalence university-affiliated individuals across diverse age cohorts. In terms occupational distribution within base, categories workers educators emerge as dominant, constituting 41% 31% demographic, respectively. Furthermore, price patrons exhibits skewed pattern, whereby bracket 0–150 encompasses 17% population, whereas range 150–300 captures notable 52%. These delineated bands collectively constitute substantial proportion, exceeding 450 embodies minority, accounting less than 20%. Notably, devised endeavor demonstrates remarkable proficiency accurately projecting forthcoming passenger volumes over seven-day horizon, significantly surpassing predictive capability logistic regression. Evidently, proffered herein excels precise anticipation counts, furnishing pragmatic foundation intelligent evolution strategies.

Язык: Английский

---

Prediction of Compressive Strength of Geopolymer Concrete Landscape Design: Application of the Novel Hybrid RF–GWO–XGBoost Algorithm

Buildings, Год журнала: 2024, Номер 14(3), С. 591 - 591

Опубликована: Фев. 22, 2024

Landscape geopolymer concrete (GePoCo) with environmentally friendly production methods not only has a stable structure but can also effectively reduce environmental damage. Nevertheless, GePoCo poses challenges its intricate cementitious matrix and vague mix design, where the components their relative amounts influence compressive strength. In response to these challenges, application of accurate applicable soft computing techniques becomes imperative for predicting strength such composite matrix. This research aimed predict using waste resources through novel ensemble ML algorithm. The dataset comprised 156 statistical samples, 15 variables were selected prediction. model employed combination RF, GWO algorithm, XGBoost. A stacking strategy was implemented by developing multiple RF models different hyperparameters, combining outcome predictions into new dataset, subsequently XGBoost model, termed RF–XGBoost model. To enhance accuracy errors, algorithm optimized hyperparameters resulting in RF–GWO–XGBoost proposed compared stand-alone models, hybrid GWO–XGBoost system. results demonstrated significant performance improvement strategies, particularly assistance exhibited better effectiveness, an RMSE 1.712 3.485, R2 0.983 0.981. contrast, (RF XGBoost) lower performance.

Язык: Английский

---

Enhancing landslide susceptibility mapping incorporating landslide typology via stacking ensemble machine learning in Three Gorges Reservoir, China

Geoscience Frontiers, Год журнала: 2024, Номер 15(4), С. 101802 - 101802

Опубликована: Янв. 29, 2024

Different types of landslides exhibit distinct relationships with environmental conditioning factors. Therefore, in regions where multiple coexist, it is required to separate landslide for susceptibility mapping (LSM). In this paper, a landslide-prone area located Chongqing Province within the middle and upper reaches Three Gorges Reservoir (TGRA), China, was selected as study area. 733 were classified into three types: reservoir-affected landslides, non-reservoir-affected rockfalls. Four inventory datasets 15 conditional factors trained by Machine Learning models (logistic regression, random forest, support vector machine), Deep (DL) model. After comparing using receiver operating characteristics (ROC), indexes acquired best performing These then used input generate final map based on Stacking method. The results revealed that DL model showed performance LSM without considering types, achieving an under curve (AUC) 0.854 testing 0.922 training. Moreover, when we separated LSM, AUC improved 0.026 0.044 Thus, paper demonstrates different can significantly improve quality maps. maps turn, be valuable tools evaluating mitigating hazards.

Язык: Английский

---

Optimized Random Forest Models for Rock Mass Classification in Tunnel Construction

Geosciences, Год журнала: 2025, Номер 15(2), С. 47 - 47

Опубликована: Фев. 2, 2025

The accurate prediction of rock mass quality ahead the tunnel face is crucial for optimizing construction strategies, enhancing safety, and reducing geological risks. This study developed three hybrid models using random forest (RF) optimized by moth-flame optimization (MFO), gray wolf optimizer (GWO), Bayesian (BO) algorithms to classify surrounding in real time during boring machine (TBM) operations. A dataset with 544 TBM tunneling samples included key parameters such as thrust force per cutter (TFC), revolutions minute (RPM), penetration rate (PR), advance (AR), revolution (PRev), field index (FPI), classification based on Rock Mass Rating (RMR) method. To address class imbalance, Borderline Synthetic Minority Over-Sampling Technique was applied. Performance assessments revealed MFO-RF model’s superior performance, training testing accuracies 0.992 0.927, respectively, predictors identified PR, AR, RPM. Additional validation 91 data sets confirmed reliability model unseen data, achieving an accuracy 0.879. graphical user interface also developed, enabling engineers technicians make instant reliable predictions, greatly supporting safe operational efficiency. These contribute valuable tools real-time, data-driven decision-making projects.

Язык: Английский

---

Developing hybrid ELM-ALO, ELM-LSO and ELM-SOA models for predicting advance rate of TBM

Transportation Geotechnics, Год журнала: 2022, Номер 36, С. 100819 - 100819

Опубликована: Июль 21, 2022

Язык: Английский

---

Six Novel Hybrid Extreme Learning Machine–Swarm Intelligence Optimization (ELM–SIO) Models for Predicting Backbreak in Open-Pit Blasting

Natural Resources Research, Год журнала: 2022, Номер 31(5), С. 3017 - 3039

Опубликована: Июнь 20, 2022

Abstract Backbreak (BB) is one of the serious adverse blasting consequences in open-pit mines, because it frequently reduces economic benefits and seriously affects safety mines. Therefore, rapid accurate prediction BB great significance to mine design other production activities. For this purpose, six different swarm intelligence optimization (SIO) algorithms were proposed optimize extreme learning machine (ELM) model for prediction, i.e., ELM-based particle (ELM–PSO), fruit fly (ELM–FOA), whale algorithm (ELM–WOA), lion (ELM–LOA), seagull (ELM–SOA) sparrow search (ELM–SSA). In total, 234 data records from operations Sungun Iran used study, including input parameters (special drilling, spacing, burden, hole length, stemming, powder factor) output parameter (i.e., BB). To evaluate predictive performance models initial models, indicators root mean square error (RMSE), Pearson correlation coefficient (R), determination (R 2 ), variance accounted (VAF), absolute (MAE) sum (SSE) training testing phases. The results show that ELM–LSO was best predict with RMSE 0.1129 ( R : 0.9991, 0.9981, VAF: 99.8135%, MAE: 0.0706 SSE: 2.0917) phase 0.2441 0.9949, 0.9891, 98.9806%, 0.1669 4.1710). Hence, ELM techniques combined SIO are an effective method BB.

Язык: Английский

---

COSMA-RF: New intelligent model based on chaos optimized slime mould algorithm and random forest for estimating the peak cutting force of conical picks

Transportation Geotechnics, Год журнала: 2022, Номер 36, С. 100806 - 100806

Опубликована: Июль 8, 2022

Язык: Английский

---

Stability prediction of hard rock pillar using support vector machine optimized by three metaheuristic algorithms

International Journal of Mining Science and Technology, Год журнала: 2023, Номер 33(8), С. 1019 - 1036

Опубликована: Июль 18, 2023

Hard rock pillar is one of the important structures in engineering design and excavation underground mines. Accurate convenient prediction stability great significance for space safety. This paper aims to develop hybrid support vector machine (SVM) models improved by three metaheuristic algorithms known as grey wolf optimizer (GWO), whale optimization algorithm (WOA) sparrow search (SSA) predicting hard stability. An integrated dataset containing 306 pillars was established generate SVM models. Five parameters including height, width, ratio width uniaxial compressive strength stress were set input parameters. Two global indices, local indices receiver operating characteristic (ROC) curve with area under ROC (AUC) utilized evaluate all models' performance. The results confirmed that SSA-SVM model best highest values indices. Nevertheless, performance unstable (AUC: 0.899) not good those stable 0.975) failed 0.990). To verify effectiveness proposed models, 5 field cases investigated a metal mine other collected from several published works. validation indicated obtained considerable accuracy, which means combination feasible approach predict

Язык: Английский

---