Compute quality/behavior index profiles over K

Computes co-ranking indices for all neighborhood sizes from 1 to K_max, for each combination of DR method and distance metric.

Usage

k_profiles(
  x,
  projections,
  K_max = NULL,
  metrics = c("Int-Euclidean", "Hausdorff", "Ichino-Yaguchi", "Wasserstein")
)

Arguments

x

An interval_data object (standardized).

projections

An idr_projections object.

K_max

Maximum neighborhood size (default nrow(x$centers) - 2).

metrics

Character vector of distance metrics.

Value

A data frame with columns Method, Metric, K, Q_TC, B_TC, Q_RE, B_RE, Q_LC, B_LC.

Examples

if (FALSE) { # \dontrun{
data(cars_mm)
x <- standardize(cars_mm)
proj <- run_idr(x)
profiles <- k_profiles(x, proj, K_max = 10)
} # }