l1_spectral {l1spectral}R Documentation

Run the l1-spectral clustering algorithm on one component

Description

This function runs the l1-spectral clustering algorithm on one component only.

Usage

l1_spectral(A, k, elements, pen, stab = TRUE)

Arguments

A

The adjacency matrix of the graph to cluster.

k

The number of clusters.

elements

The representative elements of the connected component to cluster.

pen

The penalty (to be chosen among "lasso" and "thresholdedLS").

stab

TRUE/FALSE indicating whether the representative elements should be stabilized (TRUE by default).

Value

The matrix of community indicators.

Author(s)

Camille Champion, Magali Champion

See Also

l1_spectralclustering, l1spectral.

Examples

 #########################################################
 # Performing the l1-spectral clustering on one component
 #########################################################

 # 1st: create data
 Data <- CreateDataSet(k=3, n=20, p=list(p_inside=0.1,p_outside=0.1))

 # 2nd: find the structure, the opt number of clusters and the representative elements
 Structure <- FindStructure(Data$A_hat)
 Clusters <- FindNbrClusters(A = Data$A_hat, structure = Structure)
 Elements <- FindElement(A = Data$A_hat, structure = Structure, clusters = Clusters)

 Structure_tmp <- Structure$groups[[1]] # the first component
 A_tmp <- Data$A_hat[Structure$groups[[1]],Structure$groups[[1]]]
 k <- Clusters$nbr_clusters$Component1 # number of clusters to create
 Elements_tmp <- list(score = Elements$score$Component1,
                      indices = Elements$indices$Component1)
       # the elements of the first component

 # 3rd: perform the l1-spectral clustering algorithm
 # (with stabilization, which is the most recommended setting)
 comm <- l1_spectral(A = A_tmp, k = k, elements = Elements_tmp, pen = "lasso", stab=TRUE)

[Package l1spectral version 0.99.6 Index]