GIS-based spatial prediction of landslide using road factors and random forest for Sichuan-Tibet Highway

Abstract: The objective of this work was to evaluate the use of covariate selection by expert knowledge on the performance of soil class predictive models in a complex landscape, in order to identify the best predictive model for digital soil mapping in the Southern region of Brazil. A total of 164 points were sampled in the field using the conditioned Latin hypercube, considering the covariates elevation, slope, and aspect. From the digital elevation model, environmental covariates were extracted, composing three sets, made up of: 21 covariates, covariates after the exclusion of the multicollinear ones, and covariates chosen by expert knowledge. Prediction was performed with the following models: decision tree, random forest, multiple logistic regression, and support vector machine. The accuracy of the models was evaluated by the kappa index (K), general accuracy (GA), and class accuracy. The prediction models were sensitive to the disproportionate sampling of soil classes. The best predicted map achieved a GA of 71% and K of 0.59. The use of the covariate set chosen by expert knowledge improves model performance in predicting soil classes in a complex landscape, and random forest is the best model for the spatial prediction of soil classes.

Download Full-text

A Novel GIS-Based Random Forest Machine Algorithm for the Spatial Prediction of Shallow Landslide Susceptibility

Forests ◽

10.3390/f11010118 ◽

2020 ◽

Vol 11 (1) ◽

pp. 118 ◽

Cited By ~ 6

Author(s):

Viet-Hung Dang ◽

Nhat-Duc Hoang ◽

Le-Mai-Duyen Nguyen ◽

Dieu Tien Bui ◽

Pijush Samui

Keyword(s):

Machine Learning ◽

Random Forest ◽

Landslide Susceptibility ◽

Spatial Prediction ◽

Shallow Landslide ◽

Machine Learning Algorithms ◽

Training Data ◽

Support Vector ◽

Conditioning Factors ◽

Susceptibility Modeling

This study developed and verified a new hybrid machine learning model, named random forest machine (RFM), for the spatial prediction of shallow landslides. RFM is a hybridization of two state-of-the-art machine learning algorithms, random forest classifier (RFC) and support vector machine (SVM), in which RFC is used to generate subsets from training data and SVM is used to build decision functions for these subsets. To construct and verify the hybrid RFM model, a shallow landslide database of the Lang Son area (northern Vietnam) was prepared. The database consisted of 101 shallow landslide polygons and 14 conditioning factors. The relevance of these factors for shallow landslide susceptibility modeling was assessed using the ReliefF method. Experimental results pointed out that the proposed RFM can help to achieve the desired prediction with an F1 score of roughly 0.96. The performance of the RFM was better than those of benchmark approaches, including the SVM, RFC, and logistic regression. Thus, the newly developed RFM is a promising tool to help local authorities in shallow landslide hazard mitigations.

Download Full-text

Soil type spatial prediction from Random Forest: different training datasets, transferability, accuracy and uncertainty assessment

Scientia Agricola ◽

10.1590/1678-992x-2017-0300 ◽

2019 ◽

Vol 76 (3) ◽

pp. 243-254 ◽

Cited By ~ 1

Author(s):

Diego Fernandes Terra Machado ◽

Sérgio Henrique Godinho Silva ◽

Nilton Curi ◽

Michele Duarte de Menezes

Keyword(s):

Random Forest ◽

Soil Type ◽

Spatial Prediction ◽

Uncertainty Assessment

Download Full-text

Spatial prediction of flood susceptibility using random-forest and boosted-tree models in Seoul metropolitan city, Korea

Geomatics Natural Hazards and Risk ◽

10.1080/19475705.2017.1308971 ◽

2017 ◽

Vol 8 (2) ◽

pp. 1185-1203 ◽

Cited By ~ 68

Author(s):

Sunmin Lee ◽

Jeong-Cheol Kim ◽

Hyung-Sup Jung ◽

Moung Jin Lee ◽

Saro Lee

Keyword(s):

Random Forest ◽

Spatial Prediction ◽

Metropolitan City ◽

Flood Susceptibility ◽

Tree Models ◽

Boosted Tree

Download Full-text

Spatial Prediction of the Forest Soil Water Holding Capacities in Temperate Region on a National Scale With Random Forest Models

10.22541/au.162280786.69827193/v1 ◽

2021 ◽

Author(s):

Hyunje Yang ◽

Honggeun Lim ◽

Hyung Tae Choi

Keyword(s):

Random Forest ◽

South Korea ◽

Soil Properties ◽

Water Content ◽

Spatial Prediction ◽

Random Forest Model ◽

National Scale ◽

Environmental Covariates ◽

Forest Model ◽

Water Holding

Soil water holding capacities (SWHCs) is important input factor in hydrological simulation models for sustainable water management. Forests that covered 63% of South Korea are the main source of clean water, and it is essential to estimate SWHCs on a nationwide scale for effective forest water resources management. However, there are a few studies estimating SWHCs on a nationwide scale in the temperate regions especially in South Korea. Fortunately, forest spatial big data have been collected on a national scale, and the nationwide prediction of the SWHC can be possible with this dataset. In this study, spatial prediction of forest SWHCs (saturated water content, water content at pF1.8 and 2.7) was conducted with 953 forest soil samples and forest spatial big dataset. 4 soil properties and 14 environmental covariates were used for predicting SWHCs. Simple linear regression and random forest model were compared for selecting the optimal predictive model. From the variable importance analysis, environmental covariates had as big importance as soil properties had. And prediction performance of the model with environmental covariates as the input data was higher than that of the model with soil properties. Comparing two models, the random forest model could accurately and stably predict SWHCs than the simple linear model. As a result of spatial prediction of SWHCs at the national scale through the random forest model and the forest spatial big dataset, it was confirmed that higher SWHCs were distributed along with the Baekdudaegan, the watershed-crest-line in South Korea.

Download Full-text

Classification random forest with exact conditioning for spatial prediction of categorical variables

Artificial Intelligence in Geosciences ◽

10.1016/j.aiig.2021.11.003 ◽

2021 ◽

Author(s):

Francky Fouedjio

Keyword(s):

Random Forest ◽

Spatial Prediction ◽

Categorical Variables

Download Full-text

Imputing Satellite-Derived Aerosol Optical Depth Using a Multi-Resolution Spatial Model and Random Forest for PM2.5 Prediction

Remote Sensing ◽

10.3390/rs13010126 ◽

2021 ◽

Vol 13 (1) ◽

pp. 126

Author(s):

Behzad Kianian ◽

Yang Liu ◽

Howard H. Chang

Keyword(s):

Random Forest ◽

Aerosol Optical Depth ◽

Optical Depth ◽

Ensemble Methods ◽

Predictive Performance ◽

Spatial Prediction ◽

Gap Filling ◽

Environmental Health Research ◽

Random Forest Models ◽

Pollution Exposure

A task for environmental health research is to produce complete pollution exposure maps despite limited monitoring data. Satellite-derived aerosol optical depth (AOD) is frequently used as a predictor in various models to improve PM2.5 estimation, despite significant gaps in coverage. We analyze PM2.5 and AOD from July 2011 in the contiguous United States. We examine two methods to aid in gap-filling AOD: (1) lattice kriging, a spatial statistical method adapted to handle large amounts data, and (2) random forest, a tree-based machine learning method. First, we evaluate each model’s performance in the spatial prediction of AOD, and we additionally consider ensemble methods for combining the predictors. In order to accurately assess the predictive performance of these methods, we construct spatially clustered holdouts to mimic the observed patterns of missing data. Finally, we assess whether gap-filling AOD through one of the proposed ensemble methods can improve prediction of PM2.5 in a random forest model. Our results suggest that ensemble methods of combining lattice kriging and random forest can improve AOD gap-filling. Based on summary metrics of performance, PM2.5 predictions based on random forest models were largely similar regardless of the inclusion of gap-filled AOD, but there was some variability in daily model predictions.

Download Full-text