We generated high-quality shallow landslide susceptibility maps for Bijar County, Kurdistan Province, Iran, using Random Forest (RAF), an ensemble computational intelligence method and three meta classifiers—Bagging (BA, BA-RAF), Random Subspace (RS, RS-RAF), and Rotation Forest (RF, RF-RAF). Modeling and validation were done on 111 shallow landslide locations using 20 conditioning factors tested by the Information Gain Ratio (IGR) technique. We assessed model performance with statistically based indexes, including sensitivity, specificity, accuracy, kappa, root mean square error (RMSE), and area under the receiver operatic characteristic curve (AUC). All four machine learning models that we tested yielded excellent goodness-of-fit and prediction accuracy, but the RF-RAF ensemble model (AUC = 0.936) outperformed the BA-RAF, RS-RAF (AUC = 0.907), and RAF (AUC = 0.812) models. The results also show that the Random Forest model significantly improved the predictive capability of the RAF-based classifier and, therefore, can be considered as a useful and an effective tool in regional shallow landslide susceptibility mapping.

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.

No abstract

This paper proposes and implements an early warning and monitoring system for rainfall-induced landslide (named as EWM RIL) with a case study at the Nam Dan landslide (northern Vietnam). The proposed system consists of six sensor nodes and one rainfall station that are used to sense large amounts of data in real-time such as soil moisture, pore-water pressure (PWP), movement status, and rainfall. In addition, a new flexible configuration for the wireless communication system is proposed that is capable not only to save the energy consuming but also to ensure the reliability of the system. Using wireless communication system, the sensed data were sent to the computer station for analyzing and predicting the instability of the landslide in terms of factor of safety (FoS) using the finite element seepage analysis and the limit equilibrium slope stability analysis methods. These methods are available in the SEEP/W and SLOPE/W modules of the GeoStudio software. Based on the analyzing results, the system proposed three warning levels for the landslide Early, Intermediate, and Imminent. Experiment result in the rainy season from August to September 2016 has proven the validity of the EWM RIL system. The result of this study is useful for landslide risk prevention and management in landslide prone-areas.

Shear strength is a crucial property of soils regarded as its intrinsic capacity to resist failure when forces act on the soil mass. This study proposes an advanced meta-leaner to discern the shear strength property and generate a reliable estimation of the ultimate shear strength of the soil. The proposed model is named as metaheuristic-optimized meta-ensemble learning model (MOMEM) and aims at helping geotechnical engineers accurately predict the parameter of interest. The MOMEM was established with the integration of the artificial electric field algorithm (AEFA) to dynamically blend the radial basis function neural network (RBFNN) and multivariate adaptive regression splines (MARS). In the framework of forming MOMEM, the AEFA consistently monitor the learning phases of the RBFNN and MARS in mining soil shear strength property through optimizing their controlling parameters, including neuron number, Gaussian spread, regularization coefficient, and kernel function parameter. Simultaneously, RBFNN and MARS are stacked via a linear combination method with dynamic weights optimized by the AEFA metaheuristic. The one-tail t test on 20 running times affirmed that with the greatest mean and standard deviation of RMSE (mean = 0.035 kg/cm2; Std. = 0.005 kg/cm2), MAE (mean = 0.026 kg/cm2; Std. = 0.004 kg/cm2), MAPE (mean = 7.9%; Std. = 1.72%), and R2 (mean = 0.826; Std. = 0.055), the MOMEM is significantly superior to other artificial intelligence-based methods. These analytical results indicate that MOMEM is an innovative tool for accurate calculating soil shear strength; thus, it provides geotechnical engineers with reliable figures to significantly increase soil-related engineering design.