Skip to contents

Constructs a 4D array of covariate values for counterfactual prediction, varying one variable across its empirical range while holding others at their mean.

Usage

get_scen_array(var_to_vary, scen_vals, node_names, var_names)

Arguments

var_to_vary

Character string naming the variable to vary.

scen_vals

Named list produced by get_scen_vals.

node_names

Character vector of node names (used for the first two array dimensions).

var_names

Character vector of all variable names (third dimension of the output).

Value

A four-dimensional array of dimension m x m x p x S, where S is the number of unique quantile values for the varying variable. Each slice along the fourth dimension holds one scenario. Off-diagonal entries for the varying variable take its quantile value, while all other variables are set to their mean. Diagonals are zero.

Examples

if (FALSE) { # \dontrun{
vals <- get_scen_vals(W)
scen <- get_scen_array("proximity", vals,
                       node_names = rownames(W),
                       var_names  = dimnames(W)[[3]])
} # }