spectralClustering: Spectral Clustering

In pfruan/abSNF: Using Association Signal Annotations to boost Similarity Network Fusion

Spectral Clustering

Description

Perform the famous spectral clustering algorithms. There are three variants. The default one is the third type.

Usage

spectralClustering(affinity, K, type = 3)

Arguments

affinity	Similarity matrix
K	Number of clusters
type	The variants of spectral clustering to use.

Value

A vector consisting of cluster labels of each sample.

Examples

#load data
data(data1)
data(data2)
data(weight1)
data(weight2)

#standard normalization of the datasets
data1 = standardNormalization(data1)
data2 = standardNormalization(data2)

# Calculate boosted distance matrices(here we calculate Euclidean Distance, 
Dist1 = dist2_w(as.matrix(data1),as.matrix(data1),weight1)
Dist2 = dist2_w(as.matrix(data2),as.matrix(data2),weight2)

# Next, construct similarity graphs
W1 = affinityMatrix(Dist1)
W2 = affinityMatrix(Dist2)

#
W = SNF(list(W1,W2), 20, 20)

# 
labels = spectralClustering(W, 3)

pfruan/abSNF documentation built on Sept. 16, 2022, 5:40 a.m.