Construct a Symmetric KNN graph — ConstructKnnGraph • LocalizedMarkerDetector

Constructs a symmetric KNN graph from a given feature space.

ConstructKnnGraph(
  knn = 5,
  feature_space,
  adjust_disconnection = TRUE,
  self_loop = 1
)

Arguments

knn: integer; the number of nearest neighbors to consider for constructing the graph. Default is 5.
feature_space: matrix; a cell-by-coordinate matrix (e.g., 20 principal components).
adjust_disconnection: boolean; TRUE for fully connecting disconnected components. Default is TRUE.
self_loop: integer; weight for self connections (default is 1).

Value

A list containing the following elements:

graph: matrix; symmetric KNN graph W, computed as pmax(1, (A + A^T) / 2).
adj_matrix: matrix; adjacency matrix A.
component: list; disconnected components in the graph, each component is represented as a subgraph.

Examples

feature_space <- matrix(rnorm(200), nrow = 10, ncol = 20)
result <- ConstructKnnGraph(knn = 3, feature_space = feature_space)
#> Constructing KNN graph
#> Remove  10  singleton cells.
graph <- result$graph
adj_matrix <- result$adj_matrix