Find Geodesic Neighbors Within a Point Cloud

Find Geodesic Neighbors Within a Point Cloud

Usage

geodesic.knn(
  X,
  k,
  k.graph = 5,
  graph = NULL,
  small.component = c("error", "truncate")
)

Arguments

X

Finite numeric matrix of observations, one per row.

k

Number of other vertices to return, an integer in 1:(nrow(X)-1). Self is excluded by vertex index; coincident other observations remain valid zero-distance neighbors. Ties are ordered by vertex index.

k.graph

Number of other neighbors used to construct the auxiliary symmetric kNN graph. Components are not automatically connected. Ignored when graph is supplied.

graph

Optional dgraph containing stored edge lengths, in X row order.

small.component

Error if fewer than k other vertices are reachable ("error", default), or return the available neighbors ("truncate").

Value

A list with nn.index and nn.dist matrices with k columns and effective.k, the number returned per row. Under explicit truncation, absent indices and distances are NA; unreachable vertices are never reported as neighbors.

See also

geodesic.knnx(), graph.geodesic.distances()

Examples

X <- cbind(0:3, 0)
geodesic.knn(X, k = 1, k.graph = 1) # one other vertex per row
#> $nn.index
#>      [,1]
#> [1,]    2
#> [2,]    1
#> [3,]    2
#> [4,]    3
#> 
#> $nn.dist
#>      [,1]
#> [1,]    1
#> [2,]    1
#> [3,]    1
#> [4,]    1
#> 
#> $effective.k
#> [1] 1 1 1 1
#>