Classify relative abundance samples to multi-dimensional grid points and calculate percent information retention for each scaled grid cell

mdgp_scale(x, parts, rpr_threshold, monotypic_threshold, verbose = FALSE)

Arguments

x: data frame with relative abundance of classes in columns generated with function relative_abundance_scaled_grid()
parts: integer that sets the class label precision of scaled classes. OPTIONS: 1 = 100%, 2 = 50%, 3 = 33.3%, 4 = 25%, 5 = 20%
rpr_threshold: integer setting the representativeness threshold, the minimum percentage of a mixed class in the scaled landscape to be retained in the classification scheme
monotypic_threshold: integer that sets the minimum percentage of a class to be considered monotypic and therefore retained in the scaled classification scheme
verbose: bolean; if TRUE progress will be printed in console; default = FALSE

Value

A data frame with first column containing the scaled category IDs followed by the original class columns with relative class abundance in percent. The last three columns are x_y = concatenated x and y coordinates, prc_inf_agr = percent information retained for each scaled grid cell, and class_name = the class labels of the scaled classes

Examples

# load categorical raster
r <- terra::rast(system.file("extdata/nlm_mid_geom_r3_sa0.tif", package = "landscapeScaling"))

# subset the raster to the lower 300 by 300 pixels
r_sub <- terra::crop(r,terra::ext(0,300,0,300))

# generate relative abundance for the scaled grid
rel_abund <- relative_abundance_scaled_grid(r_sub,class_field='cover',scale_factor=15)

head(rel_abund)
#>    x y          A        B        C
#> 1  0 0  4.8888889 38.22222 56.88889
#> 2 15 0 10.2222222 38.66667 51.11111
#> 3 30 0 20.0000000 42.66667 37.33333
#> 4 45 0 24.8888889 40.44444 34.66667
#> 5 60 0  0.8888889 28.00000 71.11111
#> 6 75 0 17.7777778 37.33333 44.88889

# classify relative abundance samples to multi-dimensional grid points 
# class-label precision = 33.3% (parts=3), 
# minimum representativeness = 10% and a monotypic threshold 90% 
mdgp_result <- mdgp_scale(rel_abund,parts=3,rpr_threshold=10,monotypic_threshold=90)
#> [1] "number of cells: 400"
#> [1] "number of grid points: 10"
#> [1] "number of grid points remaining: 5"

head(mdgp_result)
#>   cls      A      B      C    x_y prc_inf_agr class_name
#> 1  10  0.000  8.444 91.556  285_0      91.556       C100
#> 2  10  0.000  9.333 90.667 195_30      90.667       C100
#> 3  10  0.444 12.444 87.111 60_285      87.111       C100
#> 4   9  4.889 38.222 56.889    0_0      90.222  C67_x_B33
#> 5   9 10.222 38.667 51.111   15_0      84.444  C67_x_B33
#> 6   9 10.667 42.667 46.667  225_0      80.000  C67_x_B33