Prediction of vertical well inclination angle based on stacking ensemble learning DOI Creative Commons
Hao Yan, Shuangjin Zheng, Hongfei Chen

et al.

All Earth, Journal Year: 2024, Volume and Issue: 36(1), P. 1 - 16

Published: Nov. 27, 2024

Well deviation is a common technical challenge in vertical well drilling operations. To accurately predict the Inclination angle certain oilfield Xinjiang work area, Stacking-based ensemble learning method was established using historical data from this area. This integrates Support Vector Machine (SVM), Random Forest (RF), and K-Nearest Neighbours (KNN) algorithms through Stacking strategy. Genetic were employed to optimise parameters of each base model. The study resulted prediction for suitable oilfield. Field test results show that optimised model has best effect, with 95.3% hit rate predicting inclination ± 0.01°, higher accuracy than both single-base learners traditional models. provides new approach optimising construction oilfield, thereby improving efficiency

Language: Английский

Synthesizing Local Capacities, Multi-Source Remote Sensing and Meta-Learning to Optimize Forest Carbon Assessment in Data-Poor Regions DOI Creative Commons
Kamaldeen Mohammed, Daniel Kpienbaareh, Jinfei Wang

et al.

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(2), P. 289 - 289

Published: Jan. 15, 2025

As the climate emergency escalates, role of forests in carbon sequestration is paramount. This paper proposes a framework that integrates local capacities, multi-source remote sensing data, and meta-learning to enhance forest assessment methodologies data-scarce regions. By integrating optical radar data alongside community inventories, we applied meta-modelling approach using stacked generalization ensemble estimate above-ground (AGC). We also conducted Kruskal–Wallis test determine significant differences AGC among different tree species. The (p = 1.37 × 10−13) Dunn post-hoc analysis revealed stock potential species, with Afzelia quanzensis (x~ 12 kg/ha, P-holm-adj. 0.05) locally known species M’buta 6 5.45 10−9) exhibiting significantly higher median AGC. Our results further showed combining substantially improved prediction accuracy compared single-source data. To improve assessment, employed generalization, multiple machine learning algorithms leverage their complementary strengths address individual limitations. yielded more robust estimates than conventional methods. Notably, stacking support vector machines random achieved highest (R2 0.84, RMSE 1.36), followed by an all base learners 0.83, 1.39). Additionally, our demonstrate factors such as diversity sensitivity meta-leaners optimization can influence performance.

Language: Английский

Citations

0
Hyperspectral inversion of soil organic matter based on improved ensemble learning method DOI
Junjie Liu,

Yongsheng Hong,

Bifeng Hu

et al.

Spectrochimica Acta Part A Molecular and Biomolecular Spectroscopy, Journal Year: 2025, Volume and Issue: unknown, P. 126302 - 126302

Published: April 1, 2025

Language: Английский

Citations

0
Spatial and Temporal Variations in Soil Organic Carbon in Northwestern China via Comparisons of Different Methods DOI Creative Commons
Jinlin Li, Ning Hu, Yuxin Qi

et al.

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(3), P. 420 - 420

Published: Jan. 26, 2025

Soil organic carbon (SOC) is a crucial component for investigating cycling and global climate change. Accurate data exhibiting the temporal spatial distributions of SOC are very important determining soil sequestration potential formulating strategies. An scheme mapping to establish link between environmental factors via different methods. The Shiyang River Basin third largest inland river basin in Hexi Corridor, which has closed geographical conditions relatively independent cycle system, making it an ideal area research arid areas. In this study, 65 samples were collected 21 assessed from 2011 2021 Basin. linear regression (LR) method two machine learning methods, i.e., support vector (SVR) random forest (RF), applied estimate distribution SOC. RF slightly better than SVR because its advantages comparison classification. When latitude, slope, normalized vegetation index (NDVI) used as predictor variables, best performance shown. Compared with Harmonized World Database (HWSD), optimal improved accuracy significantly. Finally, tended increase, total increase 135.94 g/kg across whole basin. northwestern part middle decreased by 2.82% industrial activities. Minqin County increased approximately 62.77% 2021. Thus, variability increased. This study provides theoretical basis basins. addition, can also provide effective scientific suggestions projects, offer key understanding cycle, change adaptation mitigation

Language: Английский

Citations

0
Hyperspectral Inversion of Soil Organic Matter Based on Improved Ensemble Learning Method DOI
Junjie Liu,

Yongsheng Hong,

Bifeng Hu

et al.

Published: Jan. 1, 2025

Language: Английский

Citations

0
Enhanced Ensemble Learning-Based Uncertainty and Sensitivity Analysis of Ventilation Rate in a Novel Radiative Cooling Building DOI Creative Commons

Majid Mohsenpour,

Mohsen Salimi,

A. Kermani

et al.

Heliyon, Journal Year: 2024, Volume and Issue: 11(1), P. e41572 - e41572

Published: Dec. 31, 2024

Language: Английский

Citations

2
Prediction of vertical well inclination angle based on stacking ensemble learning DOI Creative Commons
Hao Yan, Shuangjin Zheng, Hongfei Chen

et al.

All Earth, Journal Year: 2024, Volume and Issue: 36(1), P. 1 - 16

Published: Nov. 27, 2024

Well deviation is a common technical challenge in vertical well drilling operations. To accurately predict the Inclination angle certain oilfield Xinjiang work area, Stacking-based ensemble learning method was established using historical data from this area. This integrates Support Vector Machine (SVM), Random Forest (RF), and K-Nearest Neighbours (KNN) algorithms through Stacking strategy. Genetic were employed to optimise parameters of each base model. The study resulted prediction for suitable oilfield. Field test results show that optimised model has best effect, with 95.3% hit rate predicting inclination ± 0.01°, higher accuracy than both single-base learners traditional models. provides new approach optimising construction oilfield, thereby improving efficiency

Language: Английский

Citations

0