Search

The main objective of this study is to propose and verify a novel ensemble methodology that could improve prediction performances of landslide susceptibility models. The proposed methodology is based on the functional tree classifier and three current state-of-the art machine learning ensemble frameworks, Bagging, AdaBoost, and MultiBoost. According to current literature, these methods have been rarely used for the modeling of rainfall-induced landslides. The corridor of the National Road 32 (Vietnam) was selected as a case study. In the first stage, the landslide inventory map with 262 landslide polygons that occurred during the last 20 years was constructed and then was randomly partitioned into a ratio of 70/30 for training and validating the models. Second, ten landslide conditioning factors were prepared such as slope, aspect, relief amplitude, topographic wetness index, topographic shape, distance to roads, distance to rivers, distance to faults, lithology, and rainfall. The model performance was assessed and compared using the receiver operating characteristic and statistical evaluation measures. Overall, the FT with Bagging model has the highest prediction capability (AUC = 0.917), followed by the FT with MultiBoost model (AUC = 0.910), the FT model (AUC = 0.898), and the FT with AdaBoost model (AUC = 0.882). Compared with those derived from popular methods such as J48 decision trees and artificial neural networks, the performance of the FT with Bagging model is better. Therefore, it can be concluded that the FT with Bagging is promising and could be used as an alternative in landslide susceptibility assessment. The result in this study is useful for land use planning and decision making in landslide prone areas.

Preparation of landslide susceptibility maps is considered as the first important step in landslide risk assessments, but these maps are accepted as an end product that can be used for land use planning. The main objective of this study is to explore some new state-of-the-art sophisticated machine learning techniques and introduce a framework for training and validation of shallow landslide susceptibility models by using the latest statistical methods. The Son La hydropower basin (Vietnam) was selected as a case study. First, a landslide inventory map was constructed using the historical landslide locations from two national projects in Vietnam. A total of 12 landslide conditioning factors were then constructed from various data sources. Landslide locations were randomly split into a ratio of 70:30 for training and validating the models. To choose the best subset of conditioning factors, predictive ability of the factors were assessed using the Information Gain Ratio with 10-fold cross-validation technique. Factors with null predictive ability were removed to optimize the models. Subsequently, five landslide models were built using support vector machines (SVM), multi-layer perceptron neural networks (MLP Neural Nets), radial basis function neural networks (RBF Neural Nets), kernel logistic regression (KLR), and logistic model trees (LMT). The resulting models were validated and compared using the receive operating characteristic (ROC), Kappa index, and several statistical evaluation measures. Additionally, Friedman and Wilcoxon signed-rank tests were applied to confirm significant statistical differences among the five machine learning models employed in this study. Overall, the MLP Neural Nets model has the highest prediction capability (90.2 %), followed by the SVM model (88.7 %) and the KLR model (87.9 %), the RBF Neural Nets model (87.1 %), and the LMT model (86.1 %). Results revealed that both the KLR and the LMT models showed promising methods for shallow lan

The main objective of this study is to investigate potential application of an integrated evidential belief function (EBF)-based fuzzy logic model for spatial prediction of rainfall-induced shallow landslides in the Lang Son city area (Vietnam). First, a landslide inventory map was constructed from various sources. Then the landslide inventory map was randomly partitioned as a ratio of 70/30 for training and validation of the models, respectively. Second, six landslide conditioning factors (slope angle, slope aspect, lithology, distance to faults, soil type, land use) were prepared and fuzzy membership values for these factors classes were estimated using the EBF. Subsequently, fuzzy operators were used to generate landslide susceptibility maps. Finally, the susceptibility maps were validated and compared using the validation dataset. The results show that the lowest prediction capability is the fuzzy SUM (76.6%). The prediction capability is almost the same for the fuzzy PRODUCT and fuzzy GAMMA models (79.6%). Compared to the frequency-ratio based fuzzy logic models, the EBF-based fuzzy logic models showed better result in both the success rate and prediction rate. The results from this study may be useful for local planner in areas prone to landslides. The modelling approach can be applied for other areas.

The main objective of this study is to assess regional landslide hazards in the Hoa Binh province of Vietnam. A landslide inventory map was constructed from various sources with data mainly for a period of 21 years from 1990 to 2010. The historic inventory of these failures shows that rainfall is the main triggering factor in this region. The probability of the occurrence of episodes of rainfall and the rainfall threshold were deduced from records of rainfall for the aforementioned period. The rainfall threshold model was generated based on daily and cumulative values of antecedent rainfall of the landslide events. The result shows that 15-day antecedent rainfall gives the best fit for the existing landslides in the inventory. The rainfall threshold model was validated using the rainfall and landslide events that occurred in 2010 that were not considered in building the threshold model. The result was used for estimating temporal probability of a landslide to occur using a Poisson probability model. Prior to this work, five landslide susceptibility maps were constructed for the study area using support vector machines, logistic regression, evidential belief functions, Bayesian-regularized neural networks, and neuro-fuzzy models. These susceptibility maps provide information on the spatial prediction probability of landslide occurrence in the area. Finally, landslide hazard maps were generated by integrating the spatial and the temporal probability of landslide. A total of 15 specific landslide hazard maps were generated considering three time periods of 1, 3, and 5 years.

The objective of this study is to investigate and compare the results of three data mining approaches, the support vector machines (SVM), decision tree (DT), and Naïve Bayes (NB) models for spatial prediction of landslide hazards in the Hoa Binh province (Vietnam). First, a landslide inventory map showing the locations of 118 landslides was constructed from various sources. The landslide inventory was then randomly partitioned into 70% for training the models and 30% for the model validation. Second, ten landslide conditioning factors were selected (i.e., slope angle, slope aspect, relief amplitude, lithology, soil type, land use, distance to roads, distance to rivers, distance to faults, and rainfall). Using these factors, landslide susceptibility indexes were calculated using SVM, DT, and NB models. Finally, landslide locations that were not used in the training phase were used to validate and compare the landslide susceptibility maps. The validation results show that the models derived using SVM have the highest prediction capability. The model derived using DT has the lowest prediction capability. Compared to the logistic regression model, the prediction capability of the SVM models is slightly better. The prediction capability of the DT and NB models is lower.