Comparisons of heuristic, general statistical and machine learning models for landslide susceptibility prediction and mapping

Huang, Faming; Cao, Zhongshan; Jianfei, Guo; Jiang, Shui‐Hua; Shu, Li; Guo, Zhengfu

doi:10.1016/j.catena.2020.104580

Cited by 289 publications

(109 citation statements)

References 72 publications

Supporting

Mentioning

107

Contrasting

Unclassified

Order By: Relevance

“…The receiver operation characteristic (ROC) curve is a graphical representation of the relationship between the sensitivity and specificity of a laboratory test over all possible diagnostic cutoff values. It reflects the corrections between the "1-Specificity" Equation (4) and "Sensitivity" (equivalent to recall) [8]. We generally use the area under the ROC curve (AUC) to reflect the total accuracy of these models.…”

Section: Training Validation and Independent Testingmentioning

confidence: 99%

“…Many methodologies have been developed for landslide susceptibility mapping (LSM) based on remote sensing and GIS [8], including heuristic, general statistical and machine learning methods. Heuristic methods usually conducted by experts mainly include the analytical hierarchy process (AHP) [9], the expert knowledge system [10], and the gray relational method [8,11]. The experience of experts and their cognition to study site could greatly influence the LSM result.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Mapping Post-Earthquake Landslide Susceptibility: A U-Net Like Approach

et al. 2020

View full text Add to dashboard Cite

A serious earthquake could trigger thousands of landslides and produce some slopes more sensitive to slide in future. Landslides could threaten human’s lives and properties, and thus mapping the post-earthquake landslide susceptibility is very valuable for a rapid response to landslide disasters in terms of relief resource allocation and posterior earthquake reconstruction. Previous researchers have proposed many methods to map landslide susceptibility but seldom considered the spatial structure information of the factors that influence a slide. In this study, we first developed a U-net like model suitable for mapping post-earthquake landslide susceptibility. The post-earthquake high spatial airborne images were used for producing a landslide inventory. Pre-earthquake Landsat TM (Thematic Mapper) images and the influencing factors such as digital elevation model (DEM), slope, aspect, multi-scale topographic position index (mTPI), lithology, fault, road network, streams network, and macroseismic intensity (MI) were prepared as the input layers of the model. Application of the model to the heavy-hit area of the destructive 2008 Wenchuan earthquake resulted in a high validation accuracy (precision 0.77, recall 0.90, F1 score 0.83, and AUC 0.90). The performance of this U-net like model was also compared with those of traditional logistic regression (LR) and support vector machine (SVM) models on both the model area and independent testing area with the former being stronger than the two traditional models. The U-net like model introduced in this paper provides us the inspiration that balancing the environmental influence of a pixel itself and its surrounding pixels to perform a better landslide susceptibility mapping (LSM) task is useful and feasible when using remote sensing and GIS technology.

show abstract

Section: Training Validation and Independent Testingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Mapping Post-Earthquake Landslide Susceptibility: A U-Net Like Approach

et al. 2020

View full text Add to dashboard Cite

show abstract

“…When dealing with a specific landslide initiation process, complicated constitutive equations for slope materials and detailed boundary conditions are imperative in corresponding mathematical calculation procedures in understanding and modelling the triggering of a landslide process. When mapping regional landslide probability by using deterministic methods, several simplification procedures, including assumptions regarding homogeneous slopes and infinite slope patterns should be considered, for analyzing the slope stability [ 12 – 14 ]. Although explicit landslide triggering process can be well understood by using the numerical modelling methods, however, it is often difficult to acquire numerous parameters of each soil property under both saturated and unsaturated conditions in a wide area, and cumbersome computational procedures for solving a series of numerical models are time-consuming.…”

Section: Introductionmentioning

confidence: 99%

Landslide susceptibility assessment using the Weight of Evidence method: A case study in Xunyang area, China

et al. 2021

View full text Add to dashboard Cite

The aim of this study is to provide a landslide susceptibility map of the Xunyang District of a mountainous terrain, at the southern part of the Qin-Ba Mountain Region, which has been highly exposed to widely distributed shallow landslides over the past few decades. The Weight of Evidence (WoE) method was adopted in this research considering both the presence of a certain landslide causative factor class and the absence of remaining classes, which was used for determining a clearly spatial correlation between a landslide occurrence and the causative factors. Intrinsic factors, including geomorphological factors, geological factors, and river flow networks, and external factors of anthropogenic engineering activities in terms of density of road network were all considered and involved in the Geological Information System (GIS) environment for reconstructing the thematic layers of factor dataset. Significant assumptions prior to the analysis were emphasized to ensure conditional independence between each pair of factors for this bivariate statistical approach. In addition, a detailed landslide inventory map was constructed through field investigation and a remote sensing interpretation process at a scale of 1:50000. The thematic layers and landslide map were overlapped to obtain a spatial statistical relationship by using the frequency ratio method. At last, the validation process for the derived susceptibility map was conducted by applying the ROC curve, indicating that more than 90% of the landslides were in categories of high and moderate susceptibility zones. The causative factor classes, including the slope angles ranging from 20 to 40°, strong weathered and fractured strata, and road network density were identified to considerably influence the landslide distribution in the study area. The results have proven to be significantly meaningful for landslide hazard risk mitigation and land use management for the local authorities responsible for these fields.

show abstract

“…In recent years, for the intrinsic capability of mining valuable information hidden in records of real slope cases, machine learning algorithms, such as artificial neural network (ANN)-based [14][15][16][17][18][19], discriminant analysis-based [20][21][22], and decision tree-based approaches [23][24][25], have been applied to the evaluation of slope stability gradually. According to the potential failure mechanism, these methods evaluate slope stability based on the characteristics of the parameters, such as geotechnical parameters, slope geometry, and water condition, and the application of these machine learning algorithms can obtain reasonable results that can be used for the evaluation of slope stability.…”

Section: Introductionmentioning

confidence: 99%

Slope stability evaluation using backpropagation neural networks and multivariate adaptive regression splines

Liao¹,

Liao

2020

Open Geosciences

View full text Add to dashboard Cite

Slope stability assessment is a critical concern in construction projects. This study explores the use of multivariate adaptive regression splines (MARS) to capture the intrinsic nonlinear and multidimensional relationship among the parameters that are associated with the evaluation of slope stability. A comparative study of machine learning solutions for slope stability assessment that relied on backpropagation neural network (BPNN) and MARS was conducted. One data set with actual slope collapse events was utilized for model development and to compare the performance of BPNN and MARS. Research results suggest that BPNN and MARS models can model the relationship between the safety factor and the slope parameters. Also, the MARS model has the advantages of computational efficiency and easy interpretation.

show abstract

Comparisons of heuristic, general statistical and machine learning models for landslide susceptibility prediction and mapping

Cited by 289 publications

References 72 publications

Mapping Post-Earthquake Landslide Susceptibility: A U-Net Like Approach

Mapping Post-Earthquake Landslide Susceptibility: A U-Net Like Approach

Landslide susceptibility assessment using the Weight of Evidence method: A case study in Xunyang area, China

Slope stability evaluation using backpropagation neural networks and multivariate adaptive regression splines

Contact Info

Product

Resources

About