Skip to contents

Generates a synthetic interval-valued dataset embedded on a Swiss roll manifold. Useful for benchmarking interval dimensionality reduction methods.

Usage

gen_interval_swissroll(
  n = 800,
  seed = 1,
  t_min = 1.5 * pi,
  t_max = 4.5 * pi,
  y_max = 70,
  noise_sd = 0.02,
  r_min = 0.2,
  r_max = 0.5
)

Arguments

n

Integer number of observations (default 800).

seed

Integer random seed (default 1).

t_min

Minimum angle parameter (default 1.5 * pi).

t_max

Maximum angle parameter (default 4.5 * pi).

y_max

Maximum value for the linear dimension (default 70).

noise_sd

Standard deviation of Gaussian noise on centers (default 0.02).

r_min

Minimum interval radius (default 0.2).

r_max

Maximum interval radius (default 0.5).

Value

An interval_data object with 3 variables (x, y, z).

Examples

dat <- gen_interval_swissroll(n = 100, seed = 42)
print(dat)
#> interval_data: 100 observations, 3 variables