The areas along transportation routes constructed in mountainous terrain often harbor significant landslide hazards. Ensemble learning techniques have proven their effectiveness in improving landslide susceptibility prediction performance. In this study, novel ensemble models (Bagging (B), Cascade Generalization (CG), and Dagging (D)) based on the Dual Perturb and Combine for Tree-based (DPCT) approach were employed to predict landslide susceptibility along the Ha Long-Van Don highway. The dataset comprised 77 landslide locations (3263 points), non-landslide locations (1:1 ratio with landslide points), and 14 conditional factors, including topography characteristics, geology, rainfall, and land use/land cover (LULC), which were input parameters for the models (B-DPCT, CG-DPCT, D-DPCT, and DPCT). Evaluation criteria for model prediction outcomes included the area under the receiver operating characteristic curve (AUC), parameters derived from the confusion matrix, the Kappa statistic, and the root mean square error (RMSE). The results demonstrate that the integration of higher-resolution datasets with hybrid machine-learning models leads to a significant improvement in predictive performance and accuracy for landslide susceptibility mapping compared to previous studies. Accordingly, landslide susceptibility maps predicted based on the B-DPCT model exhibited optimal evaluation results on the validation dataset (AUC = 0.948, accuracy ACC = 83.6, Kappa statistic = 0.67, and RMSE = 0.37), suggesting their recommended use for construction planning and mitigation efforts along the Ha Long-Van Don highway to minimize landslide-induced damages.

The mountainous terrain and monsoon-dominated climate of Central Vietnam make the region highly susceptible to rainfall-induced landslides, particularly in the Phuoc Son area, where steep slopes, complex geological conditions, and intense precipitation frequently trigger slope failures. This study develops a landslide susceptibility model using the CatBoost machine learning algorithm by integrating Landsat 8-derived surface indicators with key geo-environmental factors, including topographic, geological, hydrological, and vegetation-related parameters such as the Normalized Difference Vegetation Index (NDVI). The predictive performance of the proposed model was evaluated and compared with four widely used approaches, namely Logistic Regression (LR), Support Vector Machine (SVM), Multilayer Perceptron (MLP), and Deep Neural Network (DNN). Model accuracy was assessed using the Area Under the Receiver Operating Characteristic Curve (AUC). The results demonstrated that CatBoost outperformed all benchmark models, achieving AUCs of 0.97 and 0.93 on the training and testing datasets, respectively, indicating excellent predictive capability and strong generalization. The resulting landslide susceptibility maps effectively delineated areas with varying levels of landslide risk and provided enhanced spatial accuracy in identifying highly susceptible zones. These findings highlight the effectiveness of integrating remote sensing data with advanced ensemble machine learning techniques for landslide susceptibility assessment in tropical mountainous environments and provide a reliable scientific basis for hazard mitigation, land-use planning, and disaster risk management in Central Vietnam.

Land subsidence is one of the geotechnical phenomena that significantly affects infrastructure, construction, and sustainable urban development. Therefore, predicting and identifying areas susceptible to land subsidence has received increasing attention in recent studies. This study examines the application of deep learning models to spatial data analysis for predicting land subsidence susceptibility in the Hanoi area. Land subsidence data were obtained from satellite imagery and processed using a multi-temporal InSAR approach to determine surface deformation. They were then split into two datasets: 70% for training and 30% for validation. A total of 19 conditioning factors were used as input variables for the models, including aspect, slope, curvature, elevation, normalized difference vegetation index (NDVI), groundwater, engineering geology, hydrogeology, Holocene sediment thickness, land use/land cover (LULC), rainfall, topographic wetness index (TWI), and Landsat 8 spectral Bands 1–7. Four deep learning models, including CNN, LSTM, DRDN, and CNN-LSTM, were developed and compared to evaluate their predictive capability. The performance of the models was assessed using AUC, RMSE, MAE, and other evaluation metrics. The results show that all four models achieved good predictive performance, among which the DRDN model provided the best overall results, with AUC = 0.984, MAE = 0.064, and RMSE = 0.254 for the training dataset, while the corresponding values for the validation dataset were AUC = 0.957, MAE = 0.093, and RMSE = 0.305, indicating that the model has high accuracy and strong generalization capability in mapping land subsidence susceptibility in the study area. In addition, SHAP analysis revealed that Holocene sediment thickness and groundwater were the most important factors controlling land subsidence susceptibility in Hanoi.

Activity concentrations of 234U and 238U as well as 234U/238U ratios were investigated in the muscle and digestive diverticulum organs of ten size group samples of Asian green mussel (Perna viridis) from Binh Thuan, Vietnam by alpha-spectrometry, and radiological risk assessment was carried out. The 234U and 238U activities in muscle tissues and digestive diverticulum ranged from 1.4 ± 0.1 to 53.2 ± 4.3 Bq.kg−1, from 1.2 ± 0.2 to 57.2 ± 4.5 Bq.kg−1, and from 3.8 ± 0.3 to 45.0 ± 7.5 Bq.kg−1, from 3.1 ± 0.2 to 42.0 ± 6.6 Bq.kg−1, respectively. Uranium activities showed a decreasing trend in groups with smaller shell lengths from 5 to 6.5 cm and stabilized in individuals larger than 7 cm. Bioaccumulation factors were higher in smaller mussels, demonstrating greater accumulation, indicating their suitability as sentinel organisms for further study and radioecological monitoring of uranium isotopes. The 234U/238U activity ratios in digestive diverticula were consistently higher than in muscle tissues, suggesting dual uranium uptake pathways: direct absorption from seawater and ingestion of contaminated foods. The average risk to the mussels from exposure to uranium alone is below the ERICA screening reference level of 10 µGy.h−1, but a single sample exceeded this level indicating further investigation is necessary. Annual committed effective doses from consuming mussels containing 234U and 238U ranged from 2.11 to 43.60 µSv.y−1, with a mean of 12.04 µSv.y−1. While doses from 234 and 238U are within UNSCEAR safety thresholds, other radionuclides (e.g., Po, Pb, and Ra isotopes) should be considered for a comprehensive assessment. Not consuming small mussels is recommended due to higher potential radionuclide accumulation.

This study analyzes horizontal crustal movements across the Red River Fault Zone (RRFZ), one of the major fault systems in northern Vietnam, using GNSS data collected over nearly 30 years (1994–2023). In the ITRF2008 reference frame, GNSS stations along the fault show an average East-Southeastward motion at a rate of approximately 32.65±0.4 mm/yr, consistent with the general movement of the South China (SC) block. A slight velocity difference of about 2.5 mm/yr between stations on the SC and Sundaland (SU) blocks indicates relatively stable regional tectonic conditions, although localized deformation persists. Relative velocity analysis between the two fault flanks reveals a right-lateral strike-slip rate of approximately 2±0.5 mm/yr, accompanied by a minor extensional component of about 1±0.3 mm/yr. These findings suggest a generally stable tectonic regime for the SC block, while also implying possible contributions from subsidiary fault structures or local deformation zones.

Cadmium (Cd) is a toxic heavy metal with significant environmental and human health risks, particularly when accumulated in surface soils. Its presence reduces soil fertility, disrupts microbial ecosystems, and poses long-term ecological threats. This study explores the application of artificial intelligence (AI) models for mapping the potential distribution of Cd contamination in surface soils within the Gianh River Basin, Quang Binh Province, Vietnam. Four machine learning (ML) models Logistic Regression (LR), Radial Basis Function Network (RBFN), Random Forest (RF), and Support Vector Machine (SVM) and four deep learning (DL) model variants (DNN-Opt1 to DNN-Opt4) were developed and compared. The DNN variants differ based on the configuration of hidden layers and neuron counts. A total of 100 topsoil samples were collected and classified using the Geoaccumulation Index (Igeo), serving as the target variable for supervised learning. Thirteen conditioning factors were used as input variables, including Elevation, Soil Type, Slope, Curvature, proximity to roads and rivers, and seven Landsat 8 spectral bands. The dataset was divided into training (70%) and testing (30%) subsets. Model performance was evaluated using multiple metrics, including the area under the ROC curve (AUC), accuracy (ACC), Kappa coefficient, root mean square error (RMSE), and confusion matrix. Among the tested models, the DNN-Opt2 variant demonstrated the highest predictive performance with AUC = 0.858, ACC = 73.33%, Kappa = 0.47, and RMSE = 0.45. The resulting contamination potential map, particularly that derived from the RBFN model, categorized the region into five contamination risk levels: very low, low, moderate, high, and very high. This spatial information is critical not only for environmental management but also for assessing risks to groundwater quality and the structural integrity of buildings located in high-risk zones. The study demonstrates the efficacy of deep learning in enhancing predictive accuracy for heavy metal contamination mapping and underscores its practical relevance in civil and environmental engineering applications.

With the rapid advancement of technology, monitoring forest cover changes has become increasingly quantifiable through various techniques and methods. In this study, we developed a procedure that utilizes the Deep Neuron Network (DNN) model and the Geographic Information Systems (GIS) based on high-resolution imagery captured at different time points to create forest cover change maps in Nui Luot, Chuong My, Hanoi. Two RGB (Red-Green-Blue) spectral images were captured by Unmanned Aerial Vehicle (UAV) at two different time points (pre-scene and post-scene) and used to extract information for the DNN model to produce land cover maps for these two time points. The land cover classification was divided into four classes: (1) Trees, (2) Vacant, (3) Built area and others, and (4) Water surface. Combined with GIS analysis, the forest cover change maps were developed to quantify detailed increases or losses in forest cover based on the "Trees" class. The model's accuracy was evaluated using parameters such as the area Under the ROC Curve (AUC), Accuracy (ACC), Precision, Recall, F1-Score, Kappa, and Root Mean Square Error (RMSE). The analysis results indicate that from January 31, 2023, to October 20, 2023, the forest cover in the study area decreased by 0.53%. The accuracy metrics for the pre-change scene were: average AUC = 0.922, ACC = 76.86%, average Precision = 0.743, average Recall = 0.73, average F1-Score = 0.723, Kappa = 0.692, and RMSE = 0.297. For the post-change scene, the accuracy metrics were: average AUC = 0.954, ACC = 81.89%, average Precision = 0.823, average Recall = 0.815, average F1-Score = 0.818, Kappa = 0.758, and RMSE = 0.262. A deforestation scenario was constructed to evaluate the effectiveness of the DNN models in assessing and monitoring forest dynamics.

Advanced machine learning and deep Learning modeling applications for landslide susceptibility mapping are becoming increasingly popular. This study applied a deep learning model (DL) with a multilayer neural network to landslide research in the Phuoc Son district, Quang Nam province. Two methods for selecting conditioning factors, Correlation Attribute and OneR, were used to choose 12 condition parameters for landslides (Slope, Relief, Elevation, Distance to road, Rainfall, Land use, Weathering crust, Geology, Aspect, Soil, Distance to fault, and Curvature). Comparing the predicted results with two standard models, Naive Bayes (NB) and Support Vector Machine (SVM), showed that the DL model has higher and better prediction performance. Accordingly, the prediction performance of the DL model on the training dataset was ACC = 92.12%, AUC = 0.970, and on the validation dataset was ACC = 87.52, AUC = 0.944. The LSM developed based on the DL model indicates that areas with high landslide susceptibility are primarily concentrated in the southern part of the study area. These findings could be highly beneficial for urban planning management, risk management, and efforts to prevent and mitigate the damage caused by landslides in Phuoc Son.

In this study, we describe a comprehensive methodology to assess coastal erosion susceptibility, integrating various input factors, deep learning and machine learning models, and validation metrics. Physical and environmental variables, such as wave height and direction, magnitude of horizontal flow, geology, and slope, were used as inputs, along with coastal erosion inventories, to train and test models, including the Multi-Layer Perceptron (MLP), Functional Trees (FT), Logistic Regression (LR), Naïve Bayes (NB), Support Vector Machines (SVM), and Deep Learning (DL). The validation phase employed various metrics for assessing model performance against actual erosion inventories. Factor analysis highlighted wave direction as the most impactful variable, influencing coastal vulnerability significantly. The subsequent model performance evaluation revealed that the MLP model excelled across various criteria (e.g., sensitivity = 94.29

Studying the present strain rate is significant in determining the characteristics and origin of geological anomalies in the region. Tectonic strain occurs under the influence of various factors, especially tectonic forces, and only a few cases of deformation occur at speeds observable by humans. This research uses velocity data from GNSS measurements in Quang Nam - Quang Ngai and surrounding regions to assess present tectonic strain. The combination of methods used in this study includes calculating the ITRF Earth-fixed frame to minimize errors, the method of relative velocity calculation to compare the speed variations between station positions, and the deformation calculation method using the QOCA software developed by NASA's Jet Propulsion Laboratory (JPL). The calculated results show that the coastal areas of the study have relatively low strain rates with the principal strain rate <15 nano-strain/year, the magnitude of deformation is always less than 7.5 nano-strain/year, and the area is conducive to the development of dominant reverse faulting.

The complex iron oxide copper and gold (IOCG) Sin Quyen deposit in northern Vietnam is known as hydrothermal veins and multi-stages of mineralization. Thus, it is complicated to make a probabilistic 3D geometric model using traditional methods and to predict the hidden mineral potential. In this study, computer modeling with nearly 8000 archival data was recorded from 146 boreholes within the study area, and the chemical analysis was done on 40 samples. The 3D block model was constructed using geological structure, optimal parameters, and computational tools approach to the 3D geometric models of surface and ore bodies distribution. The Cu and Ag reserves were estimated based on the 3D geometric models. The total reserve of all ore bodies at the current depth was recorded at 540000 and 25 tons for Cu and Ag, respectively. In the study area, almost all ore bodies were observed as hydrothermal vein types, extending in Northwest-Southeast strikes and dipping around 750 m, closest to the geological observation. The mineralization characteristics of the study area are controlled by left-lateral zipper tectonic activity and faults. Based on tectonic and the 3D geometric model characteristics, the Cu ore bodies are trending continuously to more than 300m depth at the Southeast of Ngoi Phat stream, while the Northwest shows no signs.

This study compares the performance of various machine learning models for predicting landslide susceptibility using a geospatial dataset from the Lai Chau province, Vietnam. The dataset consisted of 850 landslide locations and ten influencing factors. Eight models, including Forest by Penalizing Attributes (FPA), Bagging-based FPA (BFPA), Artificial Neural Network (ANN), Logistic Regression (LR), Support Vector Machine (SVM), Multilayer Perceptron (MLP), Bayesian Network (BN), and Na & iuml;ve Bayes (NB), were evaluated based on different evaluation metrics. The results revealed distinct variations in the performance of the models across the evaluation metrics. Based on the overall rankings, the ensemble BFPA model with sensitivity=90%, specificity= 95.98%, accuracy=92.86%, Kappa=0.857, and area under the curve=0.98 demonstrated the highest capability in predicting landslide susceptibility. It was followed by BN, FPA, MLP, ANN, SVM, LR, and NB. These findings suggest that the BFPA model outperformed other models in terms of its ability to accurately identify potential landslide-prone areas in the study region. This study provides valuable insights into the comparative analysis of machine learning models for landslide susceptibility prediction. Furthermore, it supports the effectiveness of ensemble models for landslide susceptibility prediction, which can inform decision-makers, land-use planners, and disaster management agencies in making informed decisions regarding potential landslide hazards and implementing effective risk mitigation strategies in Vietnam. Continued research in this area will enhance our understanding of machine learning techniques and their application in mitigating the impact of landslides on society and the environment.

Landslides pose significant threats to lives and public infrastructure in mountainous regions. Real-time landslide monitoring presents challenges for scientists, often involving substantial costs and risks due to challenging terrain and instability. Recent technological advancements offer the potential to identify landslide-prone areas and provide timely warnings to local populations when adverse weather conditions arise. This study aims to achieve three key objectives: (1) propose indicators for detecting landslides in both field and remote sensing images; (2) develop deep learning (DL) models capable of automatically identifying landslides from fusion data of Sentinel-1 (SAR) and Sentinel-2 (optical) images; and (3) employ DL-trained models to detect this natural hazard in specific regions of Vietnam. Twenty DL models were trained, utilizing three U-shaped architectures, which include U-Net and U-Net3+, combined with different data-fusion choices. The training data consisted of multi-temporal Sentinel images and increased the accuracy of DL models using Adam optimizer to 99% in landslide detection with low loss function values. Using two bands of the Sentinel-1 could not define the characteristics of landslide traces. However, the integration between Sentinel-2 data and these bands makes the landslide detection process more effective. Therefore, the authors proposed a consolidated strategy based on three models: (1) UNet using four S2-bands, (2) UNet3+ using four S2-bands, (3) UNet using four S2-bands and VV S1-band, and (4) UNet using four S2-bands and VH S1-band for fully detect landslides. This integrated strategy uses the capabilities of each model and overcomes model result constraints to better describe landslide traces in varied geographical locations.

At approximately 5:00 AM on December 16, 2016, a rapid and deep-seated landslide was triggered by intense rainfall in the Van Hoi irrigation reservoir in Binh Dinh province, Vietnam. The landslide generated an impulsive wave with a height of approximately 20 m, resulting in severe damage to the reservoir operation station. This study investigated the mechanisms behind the landslide's initiation and simulated its initiation and motion processes through site surveys, ring shear tests, and the LS-RAPID simulation model. The physical tests were conducted on two soil samples from the sliding zone to examine the landslide mechanism. The results indicated that only sample 2 (a sand sample of completely weathered gneiss rock) showed a high level of landslide mobility due to its liquefaction phenomena resulting in a rapid pore water pressure development and a significant strength loss. In contrast, sample 1 (a silty sand sample of residual soils) did not exhibit this behavior due to its high shear resistance value at a steady state. The findings suggest that the sliding plane of the Van Hoi landslide formed in the completely weathered gneiss layer, and the high mobility level of sample 2 is primarily responsible for its rapid movement. Notably, the LS-RAPID model successfully reproduced the landslide process using the geotechnical properties obtained in the ring shear experiments. The simulation showed that the Van Hoi deep-seated landslide was initiated from the lower middle slope at a critical value of 0.55 for the pore water pressure ratio and traveled at a high velocity of approximately 37.0 m/s. The consistency between the computer simulation results and the on-site evidence and recorded data highlights the reliability of the LS-RAPID model as a tool for assessing landslide hazards.

The length of global coastline is about 356 thousand kilometers with various dynamic natural and anthropogenic. Although the number of studies on coastal landscape categorization has been increasing, it is still difficult to distinguish precisely them because the used methods commonly are traditional qualitative ones. With the leverage of remote sensing data and GIS tools, it helps categorize and identify a variety of features on land and water based on multi-source data. The aim of study is using different natural - social profile data obtained from ALOS, NOAA, and multi-temporal Landsat satellite images as input data of the convolutional-neural-network (CvNet) models for coastal landscape classification. Studies used 900 cut-line samples which represent coastal landscapes in Vietnam for training and optimizing CvNet models. As a result, nine coastal landscapes were identified including: deltas, alluvial, mature and young sand dunes, cliff, lagoon, tectonic, karst, and transitional landscapes. Three CvNet models using three different optimizer types classified the landscapes of other 1150 cut-lines in Vietnam with the accuracies about 98% and low loss function value. Excepting dalmatian, karst and delta coastal landscapes, five others distribute heterogeneous along the coasts in Vietnam. Therefore, the evaluation of additional natural components is necessary and CvNet model have ability to update new landscape types in variety of tropical nation as a step toward coastal landscape classification at both national and global scales.

The sustainability of water resource management remains challenging in many regions around the world. Yet while the significance of groundwater potential maps in water resource management is well known, no agreed-upon approach has been suggested for the production of reliable, accurate maps of groundwater potential. In this study, we evaluated the Partial Decision Tree (PART), Fuzzy Unordered Rule Induction Algorithm (FURIA), Multilayer Perception Network (MLP), Forest by Penalizing Attributes (FPA), and an ensemble version of the FPA method with the Decorate ensemble learning techniques (DFPA) for their capability to explore the associations between the locations of groundwater wells and a set of geo-environmental variables for the prediction of the potential for groundwater occurrence. We applied the methods to a spatially explicit dataset from five provinces of the Central Highlands, Vietnam. The results revealed that rainfall, land use/cover, elevation, and river density contributed most to groundwater potential in the study area. The ensemble model, i.e., DFPA, achieved greater goodness-of-fit and predictive ability than the single models. The ensemble DFPA model with accuracy = 70%, ROC-AUC = 0.77, RMSE = 0.44 provided the most accurate prediction of groundwater potential in the study area, followed by the FPA (ROC-AUC = 0.76), PART (ROC-AUC = 0.72), FURIA (ROC-AUC = 0.7), and MLP (ROC-AUC = 0.69) models, respectively. The ensemble DFPA model classified 34.7, 44.1, and 21.2% of the Central Highlands into low, moderate, and high potential categories, respectively. We experimentally showed that ensemble modeling is promising as a supporting tool in helping decision-makers, stakeholders, and researchers promote strategies for sustainable water resources management.

Globally, coastal erosion significantly impacts the socio-economic conditions and infrastructure development of coastal regions, with Vietnam facing considerable challenges due to its extensive coastline. This study focuses on developing innovative hybrid machine learning models, namely BLWL and CGLWL, which combine Locally Weighted Learning (LWL) and two optimization techniques, namely Bagging and Cascade Generalization, respectively. Quang Nam Province in Vietnam consistently affected by coastal erosions, serves as the case study. For model development, a set of historical coastal erosions and the affecting factors, such as magnitude of horizontal flow (sea currents), wave height, wave direction, distance to fault, geology, river density, elevation, curvature, aspect, slope degree, and topographic wetness index were collected and used for generation of the database. For the selection and prioritization of affecting coastal erosion factors, Correlation Attribute Evaluation (CAE) method was used. Performance of the models was evaluated using standard statistical measures: Accuracy Assessment (ACC), Sensitivity (SST), Specificity (SPF), Root Mean Squared Errors (RMSE), Kappa (K), Positive Predictive Value (PPV), and Negative Predictive Value (NPV), and Area Under the ROC Curve (AUC). Results indicated that the BLWL model (AUC: 0.978) was the best, followed by CGLWL (AUC: 0.968) and LWL (AUC: 0.963) models in accurately predicting coastal erosion susceptible areas. Therefore, it can be concluded that BLWL is a promising tool for the development of coastal erosion susceptibility maps, facilitating effective planning and management to mitigate the impact of coastal erosion.

The Red River is one of the largest rivers that plays an important role in the economic development of North Vietnam. There are many radionuclides bearing rare earth, uranium ore mines, mining industrial zones and magma intrusive formations along this river. The contamination and accumulation of radionuclides could exist at high concentration in surface sediments of this river. Thus, the present investigation aims to study the activity concentrations of 226Ra, 232Th (228Ra), 40K, and 137Cs in Red River surface sediments. Thirty sediment samples were collected, and their activity concentration was calculated using high-purity germanium gamma-ray detector. The observed results ranged from 51.0 ± 2.1 to 73.6 ± 3.7 for 226Ra, 71.4 ± 3.6 to 103 ± 5.2 for 232Th, 507 ± 24.0 to 846 ± 42.3 for 40K, and ND (not detected) to 1.33 ± 0.06 Bq/kg for 137Cs, respectively. In general, the natural radionuclides concentration of 226Ra, 232Th (228Ra), and 40K is higher than the average world average values. This indicated that the natural radionuclides could contribute from similar and principal sources surrounding the upstream of Lao Cai where distributed uranium ore mines, radionuclide bearing rare earth mines, mining industrial zones and intrusive formations. Regarding the radiological hazard assessment, results of the indices computed such as absorbed gamma dose rate (D), the excess lifetime cancer risk (ELCR), and the annual effective dose equivalent (AEDE) were nearly two times higher than world average values.

Following the India-Asia collision,continental blocks were extruded along large sinistral strike-slip faults.The longest such fault,the Ailao Shan-Red River shear zone(ASRR),separated Indochina(Sundaland)from South China.The～1000 km-long,active Red River fault(RRF)extends along the north side of the Ailao Shan and currently exhibits a combination of right-lateral slip and normal faulting.Here,after discussing Tertiary and recent deformation along and around the RRF system(slip-sense inversion,Oligocene/Quaternary offsets,Holocene slip rates,GPS measurements,earthquake mechanisms,etc.),we focus on its Plio-Quaternary extent and kinematics from SE Yunnan of China into NW Vietnam,the western Gulf of Tonkin,and farther south all the way to Sabah.New data is used to corroborate that,past the triple junction with the Dien Bien Phu fault in NW Vietnam,most of the present-day right-lateral movement between South China and Sunda blocks continues chiefly southeastwards of the Day Nui Con Voi along the Da River fault,which is roughly parallel to the RRF and was the site of the 2020,Mw 5.0 Moc Chau earthquake.We further show that this active fault likely extends much farther south along the western edge of the Oligo-Miocene Yingehai/Song Hong basin and the SE coast of Vietnam(Quy Nhon shear zone),at least to the'Ile des Cendres'volcanic alignment,and possibly farther to the western tip of the Sabah-Brunei thrust belt,offshore the active margin of northern Borneo.Finally,we discuss the kinematic consequences of large-scale tectonic inversion across much of the South China Sea,between the Philippines,Taiwan Island,and Sunda.

This study propose a new approach through which the landslide susceptibility in Quang Nam (Vietnam) will be estimated using the best model among the following algorithms: Decision Table (DT), Naïve Bayes (NB), Decision Table - Naïve Bayes (DTNB), Bagging Ensemble, Cascade Generalization Ensemble, Dagging Ensemble, Decorate Ensemble, MultiBoost Ensemble, MultiScheme Ensemble, Real Ada Boost Ensemble, Rotation Forest Ensemble, Random Sub Space Ensemble. In this regard, a map with 1130 landslide, was created and further partitioned into training (70%) and testing (30%) locations. The correlation-based features selections (CFS) method was used to select a number of 15 landslide influencing factors. Landslide locations, included in the training sample, and the landslide predictors were used as input data in order to run the above mentioned models. Kappa index, Accuracy (%) and ROC curve were employed to estimate the model’s performance and to test the outcomes provided by the models. Among the eleven machine learning algorithms, Random Sub Space Decision Table Naïve Bayes (RSSDTNB) was the most performant model with an AUC = 0.839, Accuracy = 76.55% and Kappa Index = 0.531. Therefore, this algorithm was involved in the estimation of landslide susceptibility. The Success Rate (AUC = 0.815) and Prediction Rate (AUC = 0.826) revealed the achievement of high-quality results.