CKNNConstruction: Building Unweighted Continuous K Nearest Neighbor Graph

In XuegongLab/HGC: A fast hierarchical graph-based clustering method

Description

This function builds a Continuous K Nearest Neighbor (CKNN) graph in the input feature space using Euclidean distance metric.

Usage

CKNN.Construction(mat, k, delta)

Arguments

mat	the input data saved as a numerical matrix. The columns are the features and the rows are the samples.
k	the number of nearest neighbors for building the CKNN graph.
delta	the parameter related with the distance threshold.

Details

This function fist built a KNN graph from the input data. Then the CKNN graph is built from the KNN graph. For node i and node j in the KNN graph, CKNN will link them if the distance d(i,j) between node i and node j is less than δ times of the geometric mean of d_k(i) and d_k(j). Here δ is the parameter, d_k(i) and d_k(j) are the distances from node i or node j to their k nearest neighbor.

Value

An n by n binary dgCMatrix object C, where n is the number of input samples. The matrix C is the adjacency matrix of the built CKNN graph. C[i,j] = 1 means that there is an edge between sample i and sample j.

Examples

data(Pollen)
Pollen.PCs <- Pollen[["PCs"]]
G <- CKNN.Construction(Pollen.PCs)

XuegongLab/HGC documentation built on Dec. 18, 2021, 7:23 p.m.