no0mat: Spectral Clustering

In kevinvergin/BATSLSA: R scripts for the analysis of data presented in the article "Marine bacterioplankton consortia follow deterministic, non-neutral community assembly rules" by Vergin et al.

Description

This analysis measures the similarity of exact connections for each pair of NTUs. The similarities are not weighted by phylogenetic distance. There are several options for spectral clustering kernels ( vanilladot (linear), polydot (polynomial), rbfdot (Gaussian), and others. The numbers of centers should be determined using other methods, such as the pamk function in the fpc package. This script takes a long time to run and could be improved by splitting tasks to separate processors.

Usage

no0mat(matrix, namesvec)

Arguments

inputParameter1	matrix is the matrix of connections after the diagnostic scripts are run inputParameter1
inputParameter2	iteration is the number of bootstrap replicates to run inputParameter2
inputParameter3	fn1 is a file name for the matrix of clustering results inputParameter3
inputParameter4	namesvec is a vector of names for the NTUs inputParameter4
inputParameter5	centersnum is the number of clusters expected inputParameter5
inputParameter6	kerneltype is the kernel used for spectral clustering inputParameter6
inputParameter7	matrix2 is the result from the bootstrapping function inputParameter7
inputParameter8	fn2 is the file name for the similarity matrix for the clustering results inputParameter8

Value

output 2 matrices, first the clustering results for each NTU and each iteration, and second, a #' similarity matrix based on the clustering assignments from the first part of the function

Examples

Run script with these two commands:  test1 <- bootstrap_clustering(matrix=,iteration=100,fn1="test1. #' csv",namesvec=ntunames,centersnum=7,kerneltype=rbfdot)
test2 <- clust_sim(matrix=, matrix2=test1,fn2="test2.csv")

kevinvergin/BATSLSA documentation built on May 20, 2019, 9:20 a.m.