clust.suite {metaCluster}

Determination of Suitable Clustering Algorithm for Metagenomics Data

Description

This function will give the best clustering algorithm for a given metagenomics data based on silhouette index for kmeans clustering, kmedoids clustering, fuzzy kmeans clsutering, DBSCAN clustering and hierarchical clsutering.

Usage

clust.suite(data, k, eps, minpts)

Arguments

data

Feature matrix consisting of different genomic features.Each row represents features corresponding to a particular individual or contig and each column represents different genomic features.

k

Optimum number of clusters

eps

Radius value for DBSCAN clustering

minpts

Minimum point value of DBSCAN clustering

Value

kmeans

Output of kmeans clustering

kmedoids

Output of kmedoids clustering

fkmeans

Output of fuzzy kmeans clustering

dbscan

Output of dbscan clustering

hierarchical

Output of hierarchical clustering

silhouette.kmeans

Silhouette plot of kmeans clustering

silhouette.kmedoids

Silhouette plot of kmedoids clustering

silhouette.fkmeans

Silhouette plot of fuzzy kmeans clustering

silhouette.dbscan

Silhouette plot of dbscan clustering

silhouette.hierarchical

Silhouette plot of hierarchical clustering

best.clustering.method

Best clustering algorithm based on silhouette index

silhouette.summary

Average silhouette width of each clustering algorithm

Author(s)

Dipro Sinha <[e-mail]>,Sayanti Guha Majumdar, Anu Sharma, Dwijesh Chandra Mishra

Examples

library(metaCluster)
data(metafeatures)
result <- clust.suite(metafeatures[1:200,],8,0.5,10)