K-Means Featurizer: A booster for intricate datasets

Earth Science Informatics, Journal Year: 2024, Volume and Issue: 17(2), P. 1203 - 1228

Published: Feb. 15, 2024

Language: Английский

---

Soft computing approaches for predicting boron contamination in arid sandstone groundwater

Earth Science Informatics, Journal Year: 2025, Volume and Issue: 18(2)

Published: Feb. 1, 2025

Language: Английский

---

Groundwater potential zoning using Logistics Model Trees based novel ensemble machine learning model

VIETNAM JOURNAL OF EARTH SCIENCES, Journal Year: 2024, Volume and Issue: unknown

Published: March 11, 2024

In this work, the main aim is to map potential zones of groundwater in Central Highlands (Vietnam) using a novel ensemble machine learning model, namely CG-LMT, which combination two advanced techniques, Cascade Generalization (CG) and Logistics Model Trees (LMT). For this, total 501 wells data set twelve affecting factors were gathered selected generate training testing datasets used for building validating model. Validation models was implemented utilizing various quantitative indices, including ROC curve. Results present study indicated that model performed well mapping modeling (AUC = 0.742), its predictive capability even better than single LMT 0.727). Thus, CG-LMT promising tool accurately predicting areas. addition, generated from helpful better-studying water resource management area.

Language: Английский

---

Assessing the relationship between landslide susceptibility and land cover change using machine learning

VIETNAM JOURNAL OF EARTH SCIENCES, Journal Year: 2024, Volume and Issue: unknown

Published: May 2, 2024

Landslides are natural disasters most frequent in the mountain region of Vietnam, producing critical damage to human lives and assets. Therefore, precisely identifying landslide occurrence probability within is essential supporting decision-makers or developers establishing effective strategies for reducing damage. This study aimed at developing a methodology based on machine learning, namely Xgboost (XGB), lightGBM, K-Nearest Neighbors (KNN), Bagging (BA) assessing connection land cover change susceptibility Da Lat City, Vietnam. 202 locations 13 potential drivers became input data model. Various statistical indices, root mean square error (RMSE), area under curve (AUC), absolute (MAE), were used evaluate proposed models. Our findings indicate that model was better than other models, as shown by AUC value 0.94, followed LightGBM (AUC=0.91), KNN (AUC=0.87), (AUC=0.81). In addition, urban areas increased during 2017-2023 from 25 km² 30 very high areas. approach can be applied test regions might represent necessary tool use planning reduce landslides.

Language: Английский

---

GROUNDWATER POTENTIAL ASSESSMENT IN GIA LAI PROVINCE (VIETNAM) USING MACHINE LEARNING, REMOTE SENSING AND GIS

Geographia Technica, Journal Year: 2024, Volume and Issue: 19(2/2024), P. 13 - 32

Published: May 15, 2024

Population growth, urbanization and rapid industrial development increase the demand for water resources.Groundwater is an important resource in sustainable socio-economic development.The identification of regions with probability existence groundwater necessary helping decision makers to propose effective strategies management this resource.The objective study construct maps potential groundwater, based on machine learning algorithms, namely deep neural networks (DNNs), XGBoost (XGB), CatBoost (CB), Gia Lai province Vietnam.In study, 12 conditioning factors, elevation, aspect, curvature, slope, soil type, river density, distance road, land use/land cover (LULC), Normalized Difference Vegetation Index (NDVI), Normal Built-up (NDBI), Water (NDWI), rainfall were used, along 181 inventory points, models.The proposed models evaluated using receiver operating characteristic (ROC) curve, area under curve (AUC), root-mean-square error (RMSE), mean absolute (MAE).The results showed that predictions most accurate XGB model; CB came second, DNN was performed least well.About 4,990 km² found be category very low potential; 3,045 category; 2,426 classified as moderate, 2,665 high, 2,007 high.The methodology used creating maps.This approach, can provide valuable information factors influencing assist decisionmakers or developers managing resources sustainably.It also supports territory, including tourism.This other geographic a small change input data.

Language: Английский

---

Delineation of Groundwater Potential Using the Bivariate Statistical Models and Hybridized Multi-Criteria Decision-Making Models

Water, Journal Year: 2024, Volume and Issue: 16(22), P. 3273 - 3273

Published: Nov. 14, 2024

Identifying groundwater potential zones in a basin and developing sustainable management plan is becoming more important, especially where surface water scarce. The main aim of the study to prepare maps (GWPMs) considering bivariate statistical models frequency ratio (FR), weight evidence (WoE), multi-criteria decision-making (MCDM) model Technique for Order Preference by Similarity an Ideal Solution (TOPSIS) hybridized with FR WoE. Two distance measures, Euclidean Manhattan, were used TOPSIS evaluate their effect on GWPMs. research focused Burdur Lake catchment located southwest Türkiye. In total, 74 wells high yields chosen randomly analysis, 52 (70%) training, 22 (30%) testing processes. Sixteen conditioning factors selected. area under receiver operating characteristic (AUROC) true skill statistics (TSS) utilized examine goodness-of-fit prediction accuracy approaches. TOPSIS-WoE-Manhattan WoE gave best AUROC values 0.915 0.944 training processes, respectively. TSS 0.827 0.864 obtained TOPSIS-FR-Euclidean

Language: Английский

---