cluster_similarity {Spectrum}

cluster_similarity: cluster a similarity matrix using the Ng method

Description

This function performs clustering of a similarity matrix following the method of Ng or of Melia. We recommend using the Ng method with GMM to cluster the eigenvectors instead of k-means.

Usage

cluster_similarity(A2, k = k, clusteralg = "GMM", specalg = "Ng")

Arguments

A2

Data frame or matrix: a similarity matrix

k

Numerical value: the number of clusters

clusteralg

Character value: GMM or km clustering algorithm (suggested=GMM)

specalg

Character value: Ng or Melia variant of spectral clustering (default=Ng)

Value

A numeric vector of cluster assignments

References

Ng, Andrew Y., Michael I. Jordan, and Yair Weiss. "On spectral clustering: Analysis and an algorithm." Advances in neural information processing systems. 2002.

Meila, Marina, et al. "Spectral Clustering: a Tutorial for the 2010’s." Handbook of Cluster Analysis. CRC Press, 2016. 1-23.

Examples

ng_similarity <- cluster_similarity(missl[[1]],k=8)