Extract the paths of a fitted causal model.

Returns the estimated paths of a fitted_DAG as a long data.frame, one row per path, which is a convenient starting point for custom tables and figures. Paths that are absent from the causal model are not included.

Usage

# S3 method for class 'fitted_DAG'
as.data.frame(x, row.names = NULL, optional = FALSE, ..., order_by = "default")

Arguments

x

An object of class fitted_DAG.

row.names, optional, ...

Ignored, present for consistency with the generic.

order_by

Either "default", to order the paths by their position in the causal model, "causal", to follow the paths downstream, or "strength", to order them by the size of their coefficients. Ordering by strength is not available for models that contain both binary and continuous variables, since their coefficients are not comparable, see est_DAG(). Ordering by cause is not available for cyclical models, which average() can produce when the direction of a path is not resolved.

Value

A data.frame with columns from, to, coef and se, as well as lower and upper if the model was fitted with boot larger than 0.

Examples

  d <- DAG(LS ~ BM, NL ~ BM, DD ~ NL + LS)
  d_fitted <- est_DAG(d, rhino, rhino_tree, 'lambda')
  as.data.frame(d_fitted)
#>   from to        coef         se
#> 1   BM NL  0.43034428 0.08868688
#> 2   BM LS  0.49739369 0.08934185
#> 3   NL DD  0.63125366 0.08358289
#> 4   LS DD -0.01185413 0.08233878
  as.data.frame(d_fitted, order_by = 'strength')
#>   from to        coef         se
#> 1   LS DD -0.01185413 0.08233878
#> 2   BM NL  0.43034428 0.08868688
#> 3   BM LS  0.49739369 0.08934185
#> 4   NL DD  0.63125366 0.08358289