Bootstrap for Direct LiNGAM

Usage

lingam_direct_bootstrap(
  X,
  n_sampling,
  prior_knowledge = NULL,
  apply_prior_knowledge_softly = FALSE,
  measure = "pwling",
  reg_method = "adaptive_lasso",
  lambda = "BIC",
  init_method = "ols",
  seed = NULL,
  verbose = TRUE,
  parallel = FALSE,
  n_cores = NULL,
  compute_total_effects = TRUE
)

Arguments

X: Numeric matrix (n_samples x n_features)
n_sampling: Number of bootstrap iterations
prior_knowledge: Prior knowledge matrix (NULL allowed)
apply_prior_knowledge_softly: Apply prior knowledge softly (logical)
measure: Independence measure ("pwling" or "kernel")
reg_method: Regression method ("ols", "lasso", "adaptive_lasso", "ridge")
lambda: Lambda selection ("lambda.min", "lambda.1se", "AIC", "BIC","oracle")
init_method: Method for estimating the initial weights of adaptive LASSO regression ("ols" or "ridge"). Same as the argument of the same name in lingam_direct().
seed: Random seed (NULL allowed)
verbose: Whether to display progress (logical)
parallel: Whether to use parallel processing (logical). When TRUE, each bootstrap iteration is distributed across multiple cores.
n_cores: Number of cores to use (integer, NULL allowed). When NULL, the number of cores is limited to a maximum of 2 for safety. Ignored when parallel = FALSE.
compute_total_effects: Whether to also estimate total causal effects for every variable pair on each bootstrap iteration (logical, default TRUE). For the lasso-family regression methods this roughly doubles iteration cost (an additional regression per downstream variable beyond the adjacency-matrix fit). Set to FALSE to skip it when only edge/order stability is needed (get_probabilities(), get_causal_direction_counts(), get_directed_acyclic_graph_counts(), get_causal_order_stability()); in that case get_total_causal_effects() errors if called on the result.

Value

BootstrapResult (list)

Details

When parallel = TRUE is specified, iterations are distributed across a socket cluster created by parallel::makePSOCKcluster(). The cluster is always released via on.exit(), whether the process finishes normally or an error occurs.

On iteration failures: each bootstrap iteration is run inside a tryCatch(). If an iteration errors (e.g. a resample produces near-singular columns), a warning identifying the failed iteration is issued and that iteration is excluded from the result instead of aborting the entire run. The returned BootstrapResult reflects however many iterations actually succeeded; an error is raised only if every iteration fails.

On reproducibility: During parallel execution, L'Ecuyer parallel random number streams via parallel::clusterSetRNGStream() are used. Results are reproducible given the same seed and same n_cores, but they do not numerically match the results of sequential execution (parallel = FALSE). If you need results that exactly match the sequential version, use parallel = FALSE.

Examples

LiNGAM_sample_1000 <- generate_lingam_sample_6()

# Fast example with OLS
bs <- lingam_direct_bootstrap(LiNGAM_sample_1000$data,
  n_sampling = 10L,
  reg_method = "ols",
  seed = 42
)
#> Bootstrap: 10 iterations, method=ols (sequential)
#>   iteration 1 / 10
#>   iteration 10 / 10
#> Completed in 0.1 seconds.
get_probabilities(bs)
#>      [,1] [,2] [,3] [,4] [,5] [,6]
#> [1,]  0.0  0.1  0.5  0.9  0.1  0.0
#> [2,]  0.9  0.0  0.9  0.9  0.4  0.5
#> [3,]  0.5  0.1  0.0  0.9  0.1  0.4
#> [4,]  0.1  0.1  0.1  0.0  0.1  0.1
#> [5,]  0.9  0.6  0.9  0.9  0.0  0.5
#> [6,]  1.0  0.5  0.6  0.9  0.5  0.0

# \donttest{
# With LASSO (requires glmnet)
bs_lasso <- lingam_direct_bootstrap(LiNGAM_sample_1000$data,
  n_sampling = 30L,
  seed = 42
)
#> Bootstrap: 30 iterations, method=adaptive_lasso (sequential)
#>   iteration 1 / 30
#>   iteration 10 / 30
#>   iteration 20 / 30
#>   iteration 30 / 30
#> Completed in 1.0 seconds.

# Parallel execution on 2 cores
bs_par <- lingam_direct_bootstrap(LiNGAM_sample_1000$data,
  n_sampling = 30L,
  seed = 42,
  parallel = TRUE,
  n_cores = 2L
)
#> Bootstrap: 30 iterations, method=adaptive_lasso (parallel, 2 cores)
#> Completed in 1.7 seconds.
# }