generate_randomizations_R

Base R function to either do exact enumeration or Monte Carlo random permutations,
then keep the fraction whose T^2 is below the acceptance cutoff.

Provides hardware-accelerated tools for performing rerandomization
and randomization testing in experimental research. Using a 'JAX' backend, the
package enables exact rerandomization inference even for large experiments
with hundreds of billions of possible randomizations. Key functionalities
include generating pools of acceptable rerandomizations based on covariate
balance, conducting exact randomization tests, and performing pre-analysis
evaluations to determine optimal rerandomization acceptance thresholds. The
package supports various hardware acceleration frameworks including 'CPU',
'CUDA', and 'METAL', making it versatile across accelerated computing environments. This
allows researchers to efficiently implement stringent rerandomization designs and
conduct valid inference even with large sample sizes. The package is partly based on Jerzak and Goldstein (2023) <doi:10.48550/arXiv.2310.00861>.

Connor Jerzak

fastrerandomize

Hardware-Accelerated Rerandomization for Improved Balance

Fucheng Warren Zhu

Aniket Sachin Kamat

Rebecca Goldstein

generate_randomizations_R function

<dl><dt>n_units</dt>
<dd>Integer, total number of units.</dd>
<dt>n_treated</dt>
<dd>Integer, number of units to be assigned to treatment.</dd>
<dt>X</dt>
<dd>Covariate matrix (n_units x p).</dd>
<dt>accept_prob</dt>
<dd>Numeric in [0, 1]: keep the fraction of randomizations 
that have the lowest T^2 up to this quantile.</dd>
<dt>random_type</dt>
<dd>Either "exact" or "monte_carlo".</dd>
<dt>max_draws</dt>
<dd>If `random_type="monte_carlo"`, how many permutations to sample.</dd>
<dt>batch_size</dt>
<dd>If `random_type="monte_carlo"`, how many permutations to handle per chunk.</dd></dl>

Arguments

Generate randomizations in base R, filtering by Hotelling's T^2 acceptance — generate_randomizations_R

<dl>

<dt>n_units</dt>
<dd>Integer, total number of units.</dd>


<dt>n_treated</dt>
<dd>Integer, number of units to be assigned to treatment.</dd>


<dt>X</dt>
<dd>Covariate matrix (n_units x p).</dd>


<dt>accept_prob</dt>
<dd>Numeric in [0, 1]: keep the fraction of randomizations 
that have the lowest T^2 up to this quantile.</dd>


<dt>random_type</dt>
<dd>Either "exact" or "monte_carlo".</dd>


<dt>max_draws</dt>
<dd>If `random_type="monte_carlo"`, how many permutations to sample.</dd>


<dt>batch_size</dt>
<dd>If `random_type="monte_carlo"`, how many permutations to handle per chunk.</dd>

</dl>

generate_randomizations_R: Generate randomizations in base R, filtering by Hotelling's T^2 acceptance

Description

Usage

Value

Arguments