Landslide susceptibility mapping (LSM) based on different boosting and hyperparameter optimization algorithms: A case of Wanzhou District, China

Journal of Rock Mechanics and Geotechnical Engineering, Год журнала: 2024, Номер 16(8), С. 3221 - 3232

Опубликована: Фев. 7, 2024

Boosting algorithms have been widely utilized in the development of landslide susceptibility mapping (LSM) studies. However, these possess distinct computational strategies and hyperparameters, making it challenging to propose an ideal LSM model. To investigate impact different boosting hyperparameter optimization on LSM, this study constructed a geospatial database comprising 12 conditioning factors, such as elevation, stratum, annual average rainfall. The XGBoost (XGB), LightGBM (LGBM), CatBoost (CB) were employed construct Furthermore, Bayesian (BO), particle swarm (PSO), Hyperband (HO) applied optimizing exhibited varying performances, with CB demonstrating highest precision, followed by LGBM, XGB showing poorer precision. Additionally, displayed HO outperforming PSO BO performance. HO-CB model achieved boasting accuracy 0.764, F1-score 0.777, area under curve (AUC) value 0.837 for training set, AUC 0.863 test set. was interpreted using SHapley Additive exPlanations (SHAP), revealing that slope, curvature, topographic wetness index (TWI), degree relief, elevation significantly influenced landslides area. This offers scientific reference disaster prevention research. examines utilization various Wanzhou District. It proposes HO-CB-SHAP framework effective approach accurately forecast disasters interpret models. limitations exist concerning generalizability data processing, which require further exploration subsequent

Язык: Английский

---

Landslide susceptibility mapping core-base factors and models’ performance variability: a systematic review

Natural Hazards, Год журнала: 2024, Номер 120(14), С. 12573 - 12593

Опубликована: Май 29, 2024

Язык: Английский

---

Synthetic Aperture Radar for Geosciences

Reviews of Geophysics, Год журнала: 2024, Номер 62(3)

Опубликована: Сен. 1, 2024

Abstract Synthetic Aperture Radar (SAR) has emerged as a pivotal technology in geosciences, offering unparalleled insights into Earth's surface. Indeed, its ability to provide high‐resolution, all‐weather, and day‐night imaging revolutionized our understanding of various geophysical processes. Recent advancements SAR technology, that is, developing new satellite missions, enhancing signal processing techniques, integrating machine learning algorithms, have significantly broadened the scope depth geosciences. Therefore, it is essential summarize SAR's comprehensive applications for especially emphasizing recent technologies applications. Moreover, current SAR‐related review papers primarily focused on or data techniques. Hence, integrates with features needed highlight significance addressing challenges well explore potential solving complex geoscience problems. Spurred by these requirements, this comprehensively in‐depth reviews broadly including aspects air‐sea dynamics, oceanography, geography, disaster hazard monitoring, climate change, geosciences fusion. For each applied field, scientific produced because are demonstrated combining techniques characteristics phenomena Further outlooks also explored, such other conducting interdisciplinary research offer With support deep learning, synergy will enhance capability model, simulate, forecast greater accuracy reliability.

Язык: Английский

---

Integration of GIS and Machine Learning Techniques for Mapping the Landslide-Prone Areas in the State of Goa, India

Journal of the Indian Society of Remote Sensing, Год журнала: 2023, Номер 51(7), С. 1479 - 1491

Опубликована: Июнь 6, 2023

Abstract A landslide susceptibility map (LSM) assists in reducing the danger of landslides by locating landslide-prone locations within designated area. One that are prone to India's Western Ghats which Goa is a part. This article presents LSMs prepared for state using four standard machine learning algorithms, namely Logistic Regression (LR ), Support Vector Machine (SVM), K -Nearest Neighbour (KNN), and Random Forest (RF). In order create LSMs, 78-point inventory, as well 14 conditioning factors, has been used, including slope, elevation, aspect, total curvature, plan profile yearly rainfall, Stream Power Index, Topographic Wetness distance road, depth bedrock/soil depth, soil type, lithology, land use cover. The most pertinent features models' construction have chosen Pearson correlation coefficient test method. presence shown be strongly influenced slope terrain, annual rainfall. generated were classified into five levels ranging from very low level high susceptible. prediction accuracy, precision, recall, F1-score, area under ROC (AUC-ROC), True Skill Statistics (TSS) used analyse compare created various methodologies. All these algorithms perform pretty well, evidenced overall accuracy scores 81.90% LR, 83.33% SVM, 81.94% KNN, 86.11% RF. SVM RF better approaches forecasting vulnerability research area, according TSS data. maximum AUC-ROC 86% was achieved algorithm. results performance metrics lead conclusion tree-based approach appropriate producing LSM Goa. this study indicate more areas can found Sattari, Dharbandora, Sanguem, Canacona regions

Язык: Английский

---

The Identification and Influence Factor Analysis of Landslides Using SBAS-InSAR Technique: A Case Study of Hongya Village, China

Applied Sciences, Год журнала: 2024, Номер 14(18), С. 8413 - 8413

Опубликована: Сен. 19, 2024

On 1 September 2022, a landslide in Hongya Village, Weiyuan Town, Huzhu Tu Autonomous County, Qinghai Province, caused significant casualties and economic losses. To mitigate such risks, InSAR technology is employed due to its wide coverage, all-weather operation, cost-effectiveness detecting landslides. In this study, focusing on the SBAS-InSAR Sentinel-1A satellite data from July 2021 September/October 2022 were used accurately identify areas of active landslides analyze deformation trends, combination with geological characteristics rainfall data. The results showed that strong was detected middle back maximum rate approximately -13 mm/year. surface consisted mainly Upper Pleistocene wind-deposited loess, which extremely sensitive water. closely related rainfall, increased increase rainfall. research prove ascending descending orbit based highly feasible field monitoring great practical significance for disaster prevention mitigation.

Язык: Английский

---

Risk Assessment of Multi-Hazards in Hangzhou: A Socioeconomic and Risk Mapping Approach Using the CatBoost-SHAP Model

International Journal of Disaster Risk Science, Год журнала: 2024, Номер 15(4), С. 640 - 656

Опубликована: Авг. 1, 2024

Abstract As the global push for sustainable urban development progresses, this study, set against backdrop of Hangzhou City, one China’s megacities, addressed conflict between expansion and occurrence geological hazards. Focusing on predominant hazards troubling Hangzhou—urban road collapse, land subsidence, karst collapse—we introduced a Categorical Boosting-SHapley Additive exPlanations (CatBoost-SHAP) model. This model not only demonstrates strong performance in predicting selected typical hazards, with area under curve (AUC) values reaching 0.92, 0.94, respectively, but also, through incorporation explainable SHAP, visually presents prediction process, interrelations evaluation factors, weight each factor. Additionally, study undertook multi-hazard evaluation, producing susceptibility zoning map multiple while performing tailored analysis by integrating economic population density factors Hangzhou. research enables decision makers to transcend “black box” limitations machine learning, facilitating informed making strategic resource allocation scheduling based demographic area. approach holds potential offer valuable insights cities worldwide.

Язык: Английский

---

Demystifying machine learning for predictive analytics in construction

Elsevier eBooks, Год журнала: 2025, Номер unknown, С. 463 - 486

Опубликована: Янв. 1, 2025

Язык: Английский

---

Landslide susceptibility prediction using machine learning and remote sensing: Case study in Thua Thien Hue province, Vietnam

Geological Journal, Год журнала: 2023, Номер 59(2), С. 636 - 658

Опубликована: Сен. 20, 2023

Landslides lead to widespread devastation and significant loss of life in mountainous regions around the world. Susceptibility assessments can provide critical data help decision‐makers, for example, local authorities other organizations, mitigating landslide risk, although accuracy existing studies needs be improved. This study aims assess susceptibility Thua Thien Hue province Vietnam using deep neural networks (DNNs) swarm‐based optimization algorithms, namely Adam, stochastic gradient descent (SGD), Artificial Rabbits Optimization (ARO), Tuna Swarm (TSO), Sand Cat (SCSO), Honey Badger Algorithm (HBA), Marine Predators (MPA) Particle (PSO). The locations 945 landslides occurring between 2012 2022, along with 14 conditioning factors, were used as input build DNN DNN‐hybrid models. performance proposed models was evaluated statistical indices receiver operating characteristic curve, area under curve (AUC), root mean square error, absolute error (MAE), R 2 accuracy. All had a high prediction. DNN‐MPA model highest AUC value (0.95), followed by DNN‐HBA DNN‐ARO DNN‐Adam DNN‐SGD DNN‐TSO (0.93), DNN‐PSO (0.9) finally DNN‐SCSO (0.83). High‐precision have identified that majority western region is very highly susceptible landslides. Models like aforementioned ones support decision‐makers updating large‐scale sustainable land‐use strategies.

Язык: Английский

---

A novel prediction model construction and result interpretation method for slope deformation of deep excavated expansive soil canals

Expert Systems with Applications, Год журнала: 2023, Номер 236, С. 121326 - 121326

Опубликована: Авг. 26, 2023

Язык: Английский

---

Comparing shallow landslide susceptibility maps in Northeastern Türkiye (Beşikdüzü, Trabzon): a multivariate statistical, machine learning, and physical data-based analysis

Environmental Earth Sciences, Год журнала: 2024, Номер 83(10)

Опубликована: Май 1, 2024

Abstract This study endeavors to assess and compare the efficacy of various modeling approaches, including statistical, machine learning, physical-based models, in creation shallow landslide susceptibility maps within Besikduzu district Trabzon province, situated Black Sea Region Türkiye. The inventory data, spanning from 2000 2018, was acquired through meticulous field surveys analysis Google Earth satellite imagery. Key topographic geologic input parameters, such as slope, aspect, wetness index, stream power plan profile curvature, units, were extracted a high-resolution 10 m spatial DEM (Digital Elevation Model) 1:25,000 scaled digital geology map, respectively. Additionally, soil unit weight shear strength critical for model, determined samples. To evaluate susceptibility, logistic regression, random forest, Shalstab employed chosen methods. accuracy generated by each method assessed using area under curve method, yielding impressive values 0.99 forest 0.97 regression 0.93 model. These results underscore robust performance all three methods, suggesting their applicability generating not only but also analogous areas with similar geological characteristics.

Язык: Английский

---