generate_W_base_and_sort: Generate Base Assignments and Sorts

In OptimalRerandExpDesigns: Optimal Rerandomization Experimental Designs

Description

Generates the base vectors to be used when locating the optimal rerandomization threshold

Usage

generate_W_base_and_sort(
  X,
  max_designs = 25000,
  imbalance_function = "mahal_dist",
  r = 0,
  max_max_iters = 5
)

Arguments

X	The data as an n \times p matrix.
max_designs	The maximum number of designs. Default is 25,000.
imbalance_function	A string indicating the imbalance function. Currently, "abs_sum_difference" and "mahal_dist" are the options with the latter being the default.
r	An experimental feature that adds lower imbalance vectors to the base set using the GreedyExperimentalDesign package. This controls the number of vectors to search through on each iteration.
max_max_iters	An experimental feature that adds lower imbalance vectors to the base set using the GreedyExperimentalDesign package. The maximum number of iterations to use for the greedy search.

Value

A list including all arguments plus a matrix W_base_sorted whose max_designs rows are n-length allocation vectors and the allocation vectors are in

Author(s)

Adam Kapelner

Examples

 n = 100
 p = 10
 X = matrix(rnorm(n * p), nrow = n, ncol = p)
 X = apply(X, 2, function(xj){(xj - mean(xj)) / sd(xj)})
 S = 1000
 
 W_base_obj = generate_W_base_and_sort(X, max_designs = S)
 W_base_obj

OptimalRerandExpDesigns documentation built on Jan. 28, 2021, 5:06 p.m.