transformers parameter in generate_pseudo_pop()

[m-qin]

Hello! I had a couple clarifications/questions about the transformers parameter in generate_pseudo_pop().

1. Can I only enter "pow2" and "pow3" as transformers, or can I also enter common functions like "sqrt" or a function I define beforehand like "logit"? I'm a little confused from the documentation, which says there are [only?] 2 available transformers but users may enter a customized function.
2. If the algorithm uses these transformations to improve covariate balance, are the covariates transformed in the output of generate_pseudo_pop(), which might be called matched_pop$pseudo_pop? Are matched_pop$original_corr_results and matched_pop$adjusted_corr_results evaluated on the transformed covariates? Is it recorded anywhere which transformations were used?

Thanks!

[Naeemkh]

You can use any unary function. You can define them before using them.

pow5 <- function(x) {x^5}

See the .log file for details of transformers.

[ananodehi]

Hello,
I’ve been having some difficulty applying transformations to the confounding variables in the GPS matching. Specifically, when I use the options use_cov_transform = TRUE and transformers = list("pow2", "pow3", "abs", "scale"), the results appear identical to when use_cov_transform = FALSE.
Could you please help me understand why this might be happening?
Thank you in advance for your assistance.