freekang / attributedsbm

Fits an attributed stochastic block model for networks with multiple continuous attributes

AttributedSBM Objective: This function is for fitting the attributed stochastic block model to a network with multiple continuous attributes. Last Edited: July 8, 2018 For bugs: Please contact [email protected] #Implementation details: Tested in R version 3.4.3 Dependencies:igraph and mvtnorm

Inputs:

Network: An NxN adjaceny matrix. It can be sparse

Attribute Mat: NxP vector of attributes, where p is the number of attributes you have

Prob: An indicator for wheether your attributes represent a probability of being in each of p communities. Use 0 if your attributes are not probabilities.

Outputs: A list object with entries $Comm: Node to community assignment

$SBMProb: SBM probability parameters (between community probability matrix)

$Mean: The mean of the Gaussian describing each community. This is a list object with each entry corresponding to the community of the index.

$Cov: The covariance matrix describing each community. This is a list object with each entry corresponding to the community of the index.

Example:

library('igraph')

library('mvtnorm')

source('FitAttribute.R')

Out=FitAttribute(MyNet,MyAttribute,0)