FindElement {l1spectral}

Find the representative elements of the clusters

Description

This internal function of the l1-spectral clustering algorithm finds representative elements of the clusters, that is nodes belonging to the clusters.

Usage

FindElement(A, structure, clusters, elements = NULL)

Arguments

A

The adjacency matrix

structure

Output of the function FindStructure().

clusters

Output of the function FindNbrClusters().

elements

The representative elements of the clusters (not necessary needed). If not provided, chosen using the betweeness centrality score.

Value

A list with the following elements:

• score The edge betweenness score of all nodes,

• Nodes Vector of the representative elements.

Author(s)

Camille Champion, Magali Champion

See Also

l1_spectralclustering, l1spectral.

Examples

 ######################################################
 # Finding the representative elements of the clusters
 ######################################################

 # 1st: create data (not perturbed graph)
 Data <- CreateDataSet(k=3, n=20, p=list(p_inside=0,p_outside=0))

 # 2nd: find the structure of the graph
 Structure <- FindStructure(Data$A_hat)

 # 3rd: find the optimal number of clusters (here, 3 clusters)
 Clusters <- FindNbrClusters(A = Data$A_hat, structure = Structure, k=3)

 # 4th: find the representative elements of the clusters
 Elements <- FindElement(A = Data$A_hat, structure = Structure, clusters = Clusters)
 # if elements is not provided, the representative elements of each component are chosen
 # by maximizing the edge betweenness score

 Elements <- FindElement(A = Data$A_hat, structure = Structure,
                   clusters = Clusters, elements = c(1,5,12))