Minimal IV-GMM demo data

Generates a small synthetic local IV-GMM dataset for experimenting with oracle doubly-robust IV moments. The output is a thin, test-friendly wrapper around sim_iv_data() with shorter design names.

Usage

iv_demo_data(
  n = 200,
  p = 2,
  q = 2,
  design = c("homog", "heterog", "weak"),
  sigma = 0.5,
  clip = 0.001,
  seed = NULL,
  x0 = NULL,
  inject = NULL,
  inject_col = "Y",
  idx_inject = 3L
)

Arguments

n

Integer sample size.

p

Integer feature dimension, at least 2.

q

Integer number of instruments, between 1 and 3.

design

Character; one of "homog", "heterog", or "weak".

sigma

Noise standard deviation in Y.

clip

Clipping bound in (0, 0.5) for binary instrument probabilities.

seed

Optional integer seed for reproducibility.

x0

Optional numeric length-p target point.

inject

Optional, one of NULL, "na", or "inf"; if provided, injects a single non-finite value into inject_col at idx_inject.

inject_col

Character injection target.

idx_inject

Integer row to corrupt.

Value

A list with the same structure as sim_iv_data().

Examples

demo <- iv_demo_data(n = 120, p = 2, q = 2, design = "homog", seed = 1)
str(demo$data)
#> 'data.frame':	120 obs. of  11 variables:
#>  $ X1          : num  -0.626 0.184 -0.836 1.595 0.33 ...
#>  $ X2          : num  -0.506 1.343 -0.215 -0.18 -0.1 ...
#>  $ Y           : num  1.516 -0.111 0.494 1.608 1.316 ...
#>  $ D           : int  1 0 1 0 0 1 1 1 0 0 ...
#>  $ Z1          : num  1 0 1 0 1 0 1 1 0 0 ...
#>  $ Z2          : num  1 1 0 0 0 1 1 1 0 1 ...
#>  $ piZ1_oracle : num  0.44 0.396 0.404 0.611 0.498 ...
#>  $ piZ2_oracle : num  0.489 0.445 0.453 0.658 0.547 ...
#>  $ muY_oracle  : num  0.614 0.762 0.216 3.43 1.771 ...
#>  $ muD_oracle  : num  0.417 0.373 0.375 0.629 0.489 ...
#>  $ theta_oracle: num  -0.0291 0.6707 -0.1485 1.1838 0.8003 ...
demo$theta_x0
#> [1] 1.211738