Summarize a distance benchmark

Displays the pairwise diagnostics already computed by benchmark_mdist(), without refitting distances or recomputing diagnostics.

Usage

# S3 method for class 'MDistBenchmark'
summary(object, metric = NULL, digits = 3L, ..., n = Inf, width = NULL)

Arguments

object

An object returned by benchmark_mdist().

metric

Character vector selecting diagnostics to display. NULL (the default) displays all available diagnostics. Choices include "mad", "relative_distance", "mds_congruence", "alienation", and available ARI columns such as "ari_pam". "ari" selects all available ARI columns.

digits

Number of decimal places in the displayed tables.

...

Additional arguments passed to the matrix print method.

n

Maximum number of methods to display in each table. The default, Inf, displays every successful method. The returned tibble is not limited.

width

Width of the printed output; NULL uses the default width.

Value

Invisibly returns the complete tibble with one row per unique pair of successful specifications, regardless of metric, digits, or n. Columns include the method labels, MAD, symmetric relative distance, MDS congruence, alienation, and any requested clustering ARIs. If fewer than two specifications succeed, the tibble has zero rows.

Examples

if (FALSE) { # \dontrun{
pairwise_results <- summary(benchmark_result)
summary(benchmark_result, metric = "relative_distance")
summary(benchmark_result, metric = c("alienation", "ari"))
} # }