Estimate the total causal effects between all variables at once

Estimate the total causal effects between all variables at once

Usage

estimate_all_total_effects(
  X,
  lingam_result,
  method = "adaptive_lasso",
  lambda = "BIC",
  init_method = "ols"
)

Arguments

X

Original data (n_samples x n_features)

lingam_result

Return value of lingam_direct()

method

Regression method ("ols", "lasso", "adaptive_lasso", "ridge")

lambda

Lambda selection ("lambda.min", "lambda.1se", "AIC", "BIC")

init_method

Method for estimating the initial weights of adaptive LASSO regression ("ols" or "ridge")

Value

Matrix of total causal effects (n_features x n_features). Convention: TE[i, j] is the total causal effect from variable j to variable i (j -> i). Same index convention as the adjacency matrix adjacency_matrix. The sum of direct and indirect effects.

Examples

LiNGAM_sample_1000 <- generate_lingam_sample_6()

model <- LiNGAM_sample_1000$data |>
  lingam_direct(reg_method = "ols")

LiNGAM_sample_1000$data |>
  estimate_all_total_effects(model)
#>          x0 x1        x2        x3 x4 x5
#> x0 0.000000  0  0.000000  3.033460  0  0
#> x1 2.987779  0  1.936920 21.058733  0  0
#> x2 0.000000  0  0.000000  5.992677  0  0
#> x3 0.000000  0  0.000000  0.000000  0  0
#> x4 8.004531  0 -1.128548 18.276121  0  0
#> x5 4.014974  0  0.000000 12.179395  0  0