Add bootstrap confidence intervals to all edges of a catgraph

Calls bootstrap_ci for every edge in a catgraph object and stores the lower and upper bounds as additional edge attributes (ci_lower, ci_upper, ci_conf, ci_type).

Usage

catgraph_ci(
  x,
  R = 1000L,
  conf = 0.95,
  type = c("percentile", "bca"),
  seed = NULL,
  verbose = TRUE
)

Arguments

x: A catgraph object.
R: Integer. Number of bootstrap resamples per pair. Default 1000L.
conf: Numeric. Confidence level. Default 0.95.
type: Character. "percentile" or "bca". Default "percentile".
seed: Integer or NULL. Base seed; each pair uses seed + i to ensure per-pair reproducibility without global state. Default NULL.
verbose: Logical. Print a progress counter. Default TRUE.

Value

The input catgraph object with three new edge attributes: ci_lower, ci_upper, ci_conf, ci_type.

Details

For large graphs (many variable pairs) this function can be slow because it runs R resamples per edge. Consider lowering R or running on a pruned graph (via prune_edges) to reduce computation.

Examples

df <- as.data.frame(Titanic)
df_exp <- df[rep(seq_len(nrow(df)), df$Freq), -5]
cg <- catgraph(df_exp)
# \donttest{
cg <- catgraph_ci(cg, R = 500, seed = 1)
#> 
  Bootstrap CI: edge 1 / 6
  Bootstrap CI: edge 2 / 6
  Bootstrap CI: edge 3 / 6
  Bootstrap CI: edge 4 / 6
  Bootstrap CI: edge 5 / 6
  Bootstrap CI: edge 6 / 6
igraph::E(cg$graph)$ci_lower
#> [1] 0.37158251 0.20622301 0.25051349 0.06191884 0.41857071 0.05197030
igraph::E(cg$graph)$ci_upper
#> [1] 0.4286221 0.2609237 0.3356203 0.1651256 0.4913456 0.1415683
# }