brier_bipartite_matches: brier_bipartite_matches

In rzgross/uRbanmatching:

Description

Computes all matches, then gets the brier scores for each.

Usage

brier_bipartite_matches(
  x_mat,
  cov_x,
  weight_list,
  treat_vec,
  match_method = c("with_replacement", "optimal", "greedy"),
  n_sinks = 0L,
  caliper_list = gen_caliper_list(),
  propensity_list = match_propensity_list(NULL),
  sqrt_mahal = TRUE,
  tol_val = NULL,
  design = "cross_all",
  num_folds = 5,
  match_predict_function = match_predict_xgb(),
  silent = !interactive()
)

Arguments

x_mat	Input/design matrix
cov_x	The covariance of x_mat. Can be rank_adjusted by using covariance_with_ranks.
weight_list	List of weight vectors. See 'generate_random_weights' to automatically generate a reasonable set of vectors.
treat_vec	Logical (or 1/0) vector, indicating treatment (or control).
match_method	"with_replacement", "optimal", or "greedy"
n_sinks	Number of potential matches that don't need to be matched.
caliper_list	Forces matches that are close on some metric.
propensity_list	Default NULL. List of propensity scores.
sqrt_mahal	Whether to use square root of Mahalanobis distances.
tol_val	For optimal matches, you can set a tolerance to be within optimality of, which can be zero for perfect optimality.
design	See predict_prepare
num_folds	Number of CV folds to use
match_predict_function	Function to predict treated units.
silent	Whether to suppress messages as it's running.

Value

List of matches within sink values, and brier scores for each.

rzgross/uRbanmatching documentation built on Dec. 22, 2021, 8:20 p.m.