Skip to content. | Skip to navigation

Personal tools

Geomorphology Research Group

This Slogan Viewlet registered to Chameleon Theme
« June 2017 »
June
MoTuWeThFrSaSu
1234
567891011
12131415161718
19202122232425
2627282930
You are here: Home / Tools / Landslide susceptibility assessment / R script for landslide susceptibility assessment, by Mauro Rossi

R script for landslide susceptibility assessment, by Mauro Rossi

A script for the R free software environment for statistical computing (http://www.r-project.org/) to perform, single and combined statistical assessments of landslide susceptibility, and related zonations.The script: (i) calibrates four separate landslide susceptibility (LS) models using linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), logistic regression (LR), and a neural network (NN), (ii) merges the four single models into optimal combinations using logistic regression, (iii) determines the uncertainty associated with the susceptibility estimates, adopting a bootstrap re-sampling technique, (iv) measures the fitting performance and predictive skills of the single and the combined LS models, and (v) outputs the results in tabular (text files) and graphical (Adobe® Portable Document Format, pdf file) format. The script can be customized to analyze landslide susceptibility in different areas, provided adequate landslide and environmental information is available. The script was tested using slope units as the mapping unit of reference. Other types of mapping units (e.g., unique condition unit, hydro-morphological unit, grid cell, administrative unit), can be used without changing the script significantly. The script, was prepared by Mauro Rossi (Mauro.Rossi@irpi.cnr.it).
File Zip archive LSA Package
Zip folder containing the following files: (i) the script LS_Evaluator.r, (ii) the configuration.txt general setup file, (iii) the example dataset calibration.txt, listing the dependent and independent variables used for model calibration, (iv) the example dataset validation.txt, listing the information used for model validation, and (v) the two files output_TXT.zip and output_PDF.zip with illustrative results.