Score a MISF-based geodesic-KK fit

score.misf.geodesic.kk() summarizes external coordinates against a MISF-GKK prepared object using either the exact full geodesic-KK scorer or the sparse landmark geodesic-KK scorer, depending on score_pair_mode. When a full MISF-GKK fit is supplied, the function also carries over the multiscale stage metadata and any trace tables retained by the optimizer.

Usage

score.misf.geodesic.kk(
  fit = NULL,
  coords = NULL,
  prepared = NULL,
  stiffness = 1,
  distance_floor = 1e-08,
  edge_length_epsilon = 1e-08,
  score_pair_mode = c("full", "landmark", "auto"),
  score_full_limit = 2048L,
  score_local_nbrs = 20L,
  score_landmark_count = 8L,
  return_trace = FALSE
)

Arguments

fit

Optional fit from misf.geodesic.kk().

coords

Optional coordinate matrix used when fit is omitted.

prepared

Optional MISF-GKK prepared object used when fit is omitted.

stiffness

Global stiffness constant \(K\).

distance_floor

Small positive floor used in k_ij = K / max(g_ij, distance_floor)^2.

edge_length_epsilon

Small positive stabilizer added inside each embedded edge length.

score_pair_mode

Requested scoring pair policy: "full", "landmark", or "auto".

score_full_limit

Active-set size threshold used when score_pair_mode = "auto".

score_local_nbrs

Sparse local-neighborhood size used for landmark-based scoring.

score_landmark_count

Sparse landmark count used for landmark-based scoring.

return_trace

If TRUE, attach any trace tables stored inside fit.

Value

A one-row data frame with final MISF-GKK summary fields and either full-GKK or landmark-GKK score columns, depending on the resolved scoring policy. If return_trace = TRUE, trace list-columns are attached when available.

Examples

mesh_edges <- edges.mesh(4, 4)
prepared <- prepare.misf.geodesic.kk(
  mesh_edges, n = 16, num_init = 4,
  top_level_mode = "skip", seed = 1
)
coords <- matrix(seq_len(32), nrow = 16, ncol = 2)
score <- score.misf.geodesic.kk(coords = coords, prepared = prepared)