Package 'netClust'

Model-Based Clustering of Network Data

Description

Clustering unilayer and multilayer network data by means of finite mixtures is the main utility of 'netClust'.

Details

The DESCRIPTION file:

Package:	netClust
Type:	Package
Title:	Model-Based Clustering of Network Data
Version:	1.0.1
Date:	2020-06-09
Author:	Shuchismita Sarkar [aut, cre], Volodymyr Melnykov [aut]
Maintainer:	Shuchismita Sarkar <[email protected]>
Description:	Clustering unilayer and multilayer network data by means of finite mixtures is the main utility of 'netClust'.
License:	GPL (>= 2)
Imports:	Rcpp (>= 1.0.2)
LinkingTo:	Rcpp, RcppArmadillo
RoxygenNote:	7.1.1
Encoding:	UTF-8
NeedsCompilation:	yes
Packaged:	2020-07-06 23:52:50 UTC; shuch
Depends:	R (>= 3.5.0)
Repository:	CRAN
Date/Publication:	2020-07-07 09:30:02 UTC

Index of help topics:

netClust-package        Model-Based Clustering of Network Data
netData                 Dataset: netData
netDataID               Dataset: netDataID
netEM_multilayer        Returns the EM object for multilayer network
netEM_unilayer          Returns the EM object for unilayer network

Clustering unilayer and multilayer network data by means of finite mixtures is the main utility of 'netClust'.

Author(s)

Shuchismita Sarkar [aut, cre], Volodymyr Melnykov [aut]

Maintainer: Shuchismita Sarkar <[email protected]>

References

Sarkar, S. (2019) On the use of transformations for modeling multidimensional heterogeneous data, The University of Alabama Libraries Digital Collections

Examples

data(netData) ## Read network data 
data(netDataID) ## Read original ID for network data

n <- dim(netData)[1] ## number of nodes of the network
p <- dim(netData)[4] ## number of layers of the network
K <- 2               ## number of clusters 
y <- netData

eps=0.0001
RndStrtUni= 3
RndStrtMult= 5
SmEMUni= 2
SmEMMult= 3
ItrSmEM=5
burn = 10*n
ItrMCMC= 50*n
sSigma = 1
sPsi = 1
a=0

##########################################
### Run unilayer network EM on layer 1 ###
##########################################

x <- array(0, dim = c(n,n,2))
for (i in 1:n){
  for (j in 1:n){
    x[i,j,] <- y[i,j,,1]
  }
}
  
E <- netEM_unilayer(x, K, eps, RndStrtUni, SmEMUni, ItrSmEM, burn, ItrMCMC, sSigma,a)
cat("Unilayer network", "Original ID", netDataID, "\n")
cat("Unilayer network", "Assigned ID", E$id, "\n")

##################################
### Run multilayer network EM  ###
##################################

E <- netEM_multilayer(y,K,p, eps, RndStrtMult, SmEMMult, ItrSmEM, burn, ItrMCMC, sSigma, sPsi, n, a)
cat("Multilayer network", "Original ID", netDataID, "\n")
cat("Multilayer network", "Assigned ID", E$id, "\n")

---

Dataset: netData

Description

Network data with 10 nodes and 2 layers

Usage

data("netData")

Format

The format is: num [1:10, 1:10, 1:2, 1:2] 0 0 0 0 0 0 0 0 0 0 ...

Details

Dataset demonstrating multilayer network

Source

Sarkar, S. (2020)

References

Sarkar, S. (2019) On the use of transformations for modeling multidimensional heterogeneous data, The University of Alabama Libraries Digital Collections

Examples

data(netData)
## maybe str(netData) ; plot(netData) ...

---

Dataset: netDataID

Description

ID for netData dataset

Usage

data("netDataID")

Format

A data frame with 10 observations on the following 1 variable.

netDataID

a numeric vector

Details

ID for the dataset demonstrating multilayer network

Source

Sarkar, S. (2020)

References

Sarkar, S. (2019) On the use of transformations for modeling multidimensional heterogeneous data, The University of Alabama Libraries Digital Collections

Examples

data(netDataID)
## maybe str(netDataID) ; plot(netDataID) ...

---

Returns the EM object for multilayer network

Description

Returns the EM object for multilayer network

Usage

netEM_multilayer(
  y,
  K,
  p,
  eps,
  num_rand_start,
  num_run_smallEM,
  max_itr_smallEM,
  burn,
  MCMC_itr,
  sigma_mult,
  psi_mult,
  n,
  alpha
)

Arguments

y	multiple network
K	number of clusters
p	number of layers
eps	epsilon for convergence
num_rand_start	number of random starts
num_run_smallEM	number of runs for small EM
max_itr_smallEM	maximum number of runs for small EM
burn	number of runs for burn for Metropolis Hastings
MCMC_itr	number of runs for Metropolis Hastings iterations
sigma_mult	scaling multiplier for Sigma matrix
psi_mult	scaling multiplier for Psi matrix
n	number of nodes of the network
alpha	seed provided by the user

Value

EM object

---

Returns the EM object for unilayer network

Description

Returns the EM object for unilayer network

Usage

netEM_unilayer(
  x,
  K,
  eps,
  num_rand_start,
  num_run_smallEM,
  max_itr_smallEM,
  burn,
  MCMC_itr,
  sigma_mult,
  alpha
)

Arguments

x	multiple network
K	number of clusters
eps	epsilon for convergence
num_rand_start	number of random starts
num_run_smallEM	number of runs for small EM
max_itr_smallEM	maximum number of runs for small EM
burn	number of runs for burn for Metropolis Hastings
MCMC_itr	number of runs for Metropolis Hastings iterations
sigma_mult	scaling multiplier for Sigma matrix
alpha	seed provided by the user

Value

EM object

---