Build Spatial Network via K-Nearest Neighbors — getSpatialNeighbors

Constructs a spatial adjacency matrix using K-nearest neighbors. Each point is connected to its k nearest neighbors based on Euclidean distance.

Usage

getSpatialNeighbors_KNN(
  coords,
  k = 10L,
  mutual = FALSE,
  binary = TRUE,
  verbose = FALSE
)

Arguments

coords: Numeric matrix of spatial coordinates.
k: Integer. Number of nearest neighbors. Default is 10.
mutual: Logical. If TRUE, only mutual nearest neighbors are connected (both A->B and B->A must exist). Default is FALSE.
binary: Logical. If TRUE (default), return binary adjacency matrix. If FALSE, return distance-weighted matrix.
verbose: Logical. Whether to print progress messages. Default is FALSE.

Value

Square numeric matrix of spatial adjacency weights.

Details

Uses the RANN package for efficient nearest neighbor search with KD-trees. The resulting network may be asymmetric (A is neighbor of B doesn't mean B is neighbor of A) unless mutual = TRUE.

Examples

set.seed(42)
coords <- cbind(x = runif(50), y = runif(50))
rownames(coords) <- paste0("spot_", 1:50)
W <- getSpatialNeighbors_KNN(coords, k = 6)