The goal of GCSM is to implement the generic composite similarity
measure (GCSM), described in “A generic composite measure of similarity
between geospatial variables” by Liu et al. (2020) doi:10.1016/j.ecoinf.2020.101169.
This package also provides implementations of SSIM and CMSC. Functions
are given to compute composite similarity between vectors (e.g,
gcsm), on spatial windows (e.g., gcsm_sw) or
temporal windows (e.g., gcsm_tw). They are implemented in
C++ with RcppArmadillo.
OpenMP is used to facilitate parallel computing.
You can install the released version of GCSM from CRAN with:
install.packages("GCSM")Or install the development version from GitHub with:
# install.packages("devtools")
devtools::install_github("liuyadong/GCSM")Composite similarity between vectors:
library(GCSM)
x = runif(9)
gcsm(x, x)
#> [1] 1
cmsc(x, x)
#> [1] 1
# mean shift
gcsm(x, x - 0.2, xmin = 0, xmax = 1, ymin = 0, ymax = 1)
#> [1] 0.8
cmsc(x, x - 0.2, xmin = 0, xmax = 1, ymin = 0, ymax = 1)
#> [1] 0.96
gcsm(x, x + 0.2, xmin = 0, xmax = 1, ymin = 0, ymax = 1)
#> [1] 0.8
cmsc(x, x + 0.2, xmin = 0, xmax = 1, ymin = 0, ymax = 1)
#> [1] 0.96
## dissimilarity
y = 1 - x # y is the perfect antianalog of x
gcsm(y, x)
#> [1] -1
gcsm(y, x - 0.2, xmin = 0, xmax = 1, ymin = 0, ymax = 1)
#> [1] -0.8
gcsm(y, x + 0.2, xmin = 0, xmax = 1, ymin = 0, ymax = 1)
#> [1] -0.8
# random noise
noise = rnorm(9, mean = 0, sd = 0.1)
gcsm(x, x + noise, xmin = 0, xmax = 1, ymin = 0, ymax = 1)
#> [1] 0.8438327
cmsc(x, x + noise, xmin = 0, xmax = 1, ymin = 0, ymax = 1)
#> [1] 0.9262687
## dissimilariry
gcsm(y, x + noise, xmin = 0, xmax = 1, ymin = 0, ymax = 1)
#> [1] -0.8438327Composite similarity on spatial windows:
x = matrix(runif(36), nrow = 6, ncol = 6)
gcsm_sw(x, x + 0.2, xmin = 0, xmax = 1, ymin = 0, ymax = 1, ksize = 3)
#> [,1] [,2] [,3] [,4] [,5] [,6]
#> [1,] 0.8 0.8 0.8 0.8 0.8 0.8
#> [2,] 0.8 0.8 0.8 0.8 0.8 0.8
#> [3,] 0.8 0.8 0.8 0.8 0.8 0.8
#> [4,] 0.8 0.8 0.8 0.8 0.8 0.8
#> [5,] 0.8 0.8 0.8 0.8 0.8 0.8
#> [6,] 0.8 0.8 0.8 0.8 0.8 0.8
cmsc_sw(x, x + 0.2, xmin = 0, xmax = 1, ymin = 0, ymax = 1, ksize = 3)
#> [,1] [,2] [,3] [,4] [,5] [,6]
#> [1,] 0.96 0.96 0.96 0.96 0.96 0.96
#> [2,] 0.96 0.96 0.96 0.96 0.96 0.96
#> [3,] 0.96 0.96 0.96 0.96 0.96 0.96
#> [4,] 0.96 0.96 0.96 0.96 0.96 0.96
#> [5,] 0.96 0.96 0.96 0.96 0.96 0.96
#> [6,] 0.96 0.96 0.96 0.96 0.96 0.96
ssim_sw(x, x + 0.2, xmin = 0, xmax = 1, ymin = 0, ymax = 1, ksize = 3)
#> [,1] [,2] [,3] [,4] [,5] [,6]
#> [1,] 0.9400044 0.9439833 0.9348286 0.9220044 0.9163602 0.9286758
#> [2,] 0.9506770 0.9548234 0.9387744 0.9230389 0.8870794 0.9021104
#> [3,] 0.9493531 0.9532574 0.9279886 0.9208246 0.8784030 0.9029298
#> [4,] 0.9132100 0.9333228 0.9168807 0.8992412 0.8577266 0.8920520
#> [5,] 0.8585278 0.9051925 0.9263586 0.9268652 0.9316172 0.9424565
#> [6,] 0.8440742 0.9072937 0.9442298 0.9433688 0.9478457 0.9510028Composite similarity on temporal windows:
x = array(runif(81), dim = c(3, 3, 9))
gcsm_tw(x, x + 0.2, xmin = 0, xmax = 1, ymin = 0, ymax = 1)
#> [,1] [,2] [,3]
#> [1,] 0.8 0.8 0.8
#> [2,] 0.8 0.8 0.8
#> [3,] 0.8 0.8 0.8
cmsc_tw(x, x + 0.2, xmin = 0, xmax = 1, ymin = 0, ymax = 1)
#> [,1] [,2] [,3]
#> [1,] 0.96 0.96 0.96
#> [2,] 0.96 0.96 0.96
#> [3,] 0.96 0.96 0.96