Input data specifications. As shown by Figure 1, the software is designed to use different mapping units, but reducible to point-like units (pixels) or to polygon-like subdivisions (e.g. geomorphological, administrative, etc.). The software cannot use as input rasters. Hence, a workaround is needed to perform a pixel based analysis. This consists in exporting the pixels of a raster in a list format. This operation can be performed in different GIS environment, two possible functions are the “gdal2xyz” function (▇▇▇▇://▇▇▇.▇▇▇▇▇.▇▇▇/gdal/trunk/gdal/swig/python/scripts/gdal2xyz.py) integrated in different GIS clients (e.g. QGIS, ▇▇▇▇://▇▇▇.▇▇▇▇.▇▇▇/en/site/) or the “raster2xyz” function in the ArcGIS platform (▇▇▇▇://▇▇▇.▇▇▇▇.▇▇▇/software/arcgis). In the basic mode the software needs the training.txt and the validation.txt files (see Table 1). These two files contain the information on the mapping units for the training and validation phase respectively. The two files are tab-separated .txt file, with named columns (without spaces) containing in order: (i) the identification for the mapping units, (ii) the binary grouping variable (i.e. dependent variable) showing the absence or presence (respectively 0 and 1) of a landslide in the mapping unit, and (iii) and a set of n explanatory variables (i.e. independent variables). Depending of the user main objective, the subdivision of the training and validation dataset (Figure 1) and hence the type of the validation performed by the model, can be done in different way: temporal, spatial, random. In the case of a temporal subdivision, the two dataset files will contain the same mapping units, with the same number, the same identifications and the same values of the explanatory variables, but with different values of the grouping variable obtained using a different landslide inventory (possibly successive to that used in the training phase). In the spatial and random subdivisions, the two dataset files will contain different mapping units, with different identification, grouping variable explanatory variable values. The main difference between the two subdivision methods consist in the different method to subdivide the training and the validation datasets: in the spatial case the dataset are relative to two areas separated spatially (contiguous or not), while in the random case the subdivision is based on a random sampling method. In the advanced mode (or geo mode) the software needs two additional files: trainin.shp and validation.shp (see Table 1). These can be point or polygon shapefiles containing geographical data for each mapping unit in the training.txt and validation.txt file. The two shapefile attribute table must contains the following fields (i) the identification values, (ii) the area, and (ii) the landslide area for each mapping units reported in the two corresponding .txt files. SusceptibilityAnalysis_vX_YYYY MMDD.R R script file containing the susceptibility analysis source code configuration_spatial_data.txt File containing the parameters for the spatial data configuration configuration.txt File containing the parameters for the susceptibility model configuration training.txt Tab-separated textual input file with named columns (without spaces) containing in order (1) an ID for the mapping units, (2) the grouping variable with 0 or 1 values, (3 to N) explanatory numerical variables. Rows contain these values for the each mapping unit of the training dataset validation.txt Tab-separated textual input file with named columns (without spaces) containing in order (1) an ID for the mapping units, (2) the grouping variable with 0 or 1 values, (3 to N) explanatory numerical variables. Rows contain these values for the each mapping unit of the validation dataset training.shp Points or polygons shapefile containing the geographical data for each mapping unit in the training.txt file. The shapefile attribute table must contains the following fields (1) the ID, (2) the area, and (3) the landslide area of each mapping units (needed only in the advanced or geo mode for the spatial data output restitution and for the success rate calculation) validation.shp Points or polygons shapefile containing the geographical data for each mapping unit in the validation.txt file. The shapefile attribute table must contains the following fields (1) the ID, (2) the area, and (3) the landslide area of each mapping units (needed only in the advanced or geo mode for the spatial data output restitution and for the prediction rate calculation)
Appears in 1 contract
Sources: Grant Agreement
Input data specifications. As shown by Figure 15, the software is designed to use different mapping units, but reducible to point-like point-‐like units (pixels) or to polygon-like polygon-‐like subdivisions (e.g. geomorphological, administrative, etc.). The software cannot use as input rasters. Hence, a workaround is needed to perform a pixel based analysis. This consists in exporting the pixels of a raster in a list format. This operation can be performed in different GIS environment, two possible functions are the “gdal2xyz” function (▇▇▇▇://▇▇▇.▇▇▇▇▇.▇▇▇/gdal/trunk/gdal/swig/python/scripts/gdal2xyz.py) integrated in different GIS clients (e.g. QGIS, ▇▇▇▇://▇▇▇.▇▇▇▇.▇▇▇/en/site/) or the “raster2xyz” function in the ArcGIS platform (▇▇▇▇://▇▇▇.▇▇▇▇.▇▇▇/software/arcgis). In the basic mode the software needs the training.txt and the validation.txt files (see Table 15). These two files contain the information on the mapping units for the training and validation phase respectively. The two files are tab-separated tab-‐separated .txt file, with named columns (without spaces) containing in order: (i) the identification for the mapping units, (ii) the binary grouping variable (i.e. dependent variable) showing the absence or presence (respectively 0 and 1) of a landslide in the mapping unit, and (iii) and a set of n explanatory variables (i.e. independent variables). Depending of the user main objective, the subdivision of the training and validation dataset (Figure 15) and hence the type of the validation performed by the model, can be done in different way: temporal, spatial, random. In the case of a temporal subdivision, the two dataset files will contain the same mapping units, with the same number, the same identifications and the same values of the explanatory variables, but with different values of the grouping variable obtained using a different landslide inventory (possibly successive to that used in the training phase). In the spatial and random subdivisions, the two dataset files will contain different mapping units, with different identification, grouping variable explanatory variable values. The main difference between the two subdivision methods consist in the different method to subdivide the training and the validation datasets: in the spatial case the dataset are relative to two areas separated spatially (contiguous or not), while in the random case the subdivision is based on a random sampling method. In the advanced mode (or geo mode) the software needs two additional files: trainin.shp and validation.shp (see Table 15). These can be point or polygon shapefiles containing geographical data for each mapping unit in the training.txt and validation.txt file. The two shapefile attribute table must contains the following fields (i) the identification values, (ii) the area, and (ii) the landslide area of the landlides, for each mapping units unit, reported in the two corresponding .txt files. SusceptibilityAnalysis_vX_YYYY MMDD.R configuration_spatial_data.txt configuration.txt training.txt validation.txt training.shp validation.shp R script file containing the susceptibility analysis source code configuration_spatial_data.txt File containing the parameters for the spatial data configuration configuration.txt File containing the parameters for the susceptibility model configuration training.txt Tab-separated Tab-‐separated textual input file with named columns (without spaces) containing in order (1) an ID for the mapping units, (2) the grouping variable with 0 or 1 values, (3 to N) explanatory numerical variables. Rows contain these values for the each mapping unit of the training dataset validation.txt Tab-separated Tab-‐separated textual input file with named columns (without spaces) containing in order (1) an ID for the mapping units, (2) the grouping variable with 0 or 1 values, (3 to N) explanatory numerical variables. Rows contain these values for the each mapping unit of the validation dataset training.shp Points or polygons shapefile containing the geographical data for each mapping unit in the training.txt file. The shapefile attribute table must contains the following fields (1) the ID, (2) the area, and (3) the landslide area of each mapping units (needed only in the advanced or geo mode for the spatial data output restitution and for the success rate calculation) validation.shp Points or polygons shapefile containing the geographical data for each mapping unit in the validation.txt file. The shapefile attribute table must contains the following fields (1) the ID, (2) the area, and (3) the landslide area of each mapping units (needed only in the advanced or geo mode for the spatial data output restitution and for the prediction rate calculation)
Appears in 1 contract