FindStructure {l1spectral}

Find the structure of the graph from the adjacency matrix

Description

This internal function of the spectral clustering algorithm finds the structure of the graph to cluster (number of nodes and connected components).

Usage

FindStructure(A)

Arguments

A

The adjacency matrix

Value

A list with the following elements:

• graph igraph object derived from A,

• groups List of connected components and corresponding nodes.

Author(s)

Camille Champion, Magali Champion

See Also

l1_spectralclustering, l1spectral.

Examples

 ###############################################################
 # Finding the structure of the graph from the adjacency matrix
 ###############################################################

 # 1st example: non-perturbed graph
 Data <- CreateDataSet(k=3, n=20, p=list(p_inside=0,p_outside=0))

 Structure <- FindStructure(Data$A_hat)
 Structure$groups # the graph is not perturbed, there are 3 connected components

 # 2nd example: highly-perturbed graph
 Data <- CreateDataSet(k=3, n=20, p=list(p_inside=0.5,p_outside=0.5))

 Structure <- FindStructure(Data$A_hat)
 Structure$groups # the graph is higlhy perturbed, there are less than 3 connected components