Calculate Several Complexity Measures

Calculates the complexity measures of complexity_measures in one call.

Usage

complexity(
  x,
  y,
  measures = NULL,
  k_R = 7,
  theta = 3,
  k_type = 5,
  scale = "minmax",
  seed = NULL
)

Arguments

x

feature matrix or data.frame (numeric).

y

a factor class variable.

measures

character vector with the names of the measures, for example c("IR", "R_aug", "N3"). The default NULL calculates all of them ("IR", "C1", "C2", "ID", "F1", "F1v", "F2", "F3", "F4", "N1", "N2", "N3", "N3_bal", "N4", "LSC", "R", "R_aug", "SAFE", "BORDERLINE", "RARE", "OUTLIER").

k_R

number of neighbors of R and R_aug. Default is 7.

theta

threshold of R and R_aug. Default is 3.

k_type

number of neighbors of the types of minority examples. Default is 5.

scale, seed

see complexity_measures.

Value

a named numeric vector.

Details

The distance matrix is calculated once and shared by the measures that need it, so asking for several measures at once is faster than calling them one by one. The measures that use all pairwise distances (N1, N2, LSC) need memory of the order of \(n^2\).

References

see complexity_measures.

Author

Fatih Saglam, saglamf89@gmail.com

Examples

set.seed(1)
x <- rbind(matrix(rnorm(400, 3, 1), ncol = 2), matrix(rnorm(40, 4, 1), ncol = 2))
y <- factor(rep(c("negative", "positive"), c(200, 20)))
complexity(x, y)
#>          IR          C1          C2          ID          F1         F1v 
#> 10.00000000  0.56050301  0.80198020  0.81818182  0.58695972  0.86061593 
#>          F2          F3          F4          N1          N2          N3 
#>  0.36475028  0.69545455  0.59545455  0.18636364  0.22712955  0.11818182 
#>      N3_bal          N4         LSC           R       R_aug        SAFE 
#>  0.40250000  0.07727273  0.91997934  0.09090909  0.90909091  0.00000000 
#>  BORDERLINE        RARE     OUTLIER 
#>  0.15000000  0.50000000  0.35000000 
complexity(x, y, measures = c("IR", "R", "R_aug", "N4"), seed = 1)
#>          IR           R       R_aug          N4 
#> 10.00000000  0.09090909  0.90909091  0.12272727