Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!
Get Started for Free

edgeset.constructors

Edgeset Constructors for Network Objects


Description

These functions convert relational data in matrix form to network edge sets.

Usage

network.bipartite(x, g, ignore.eval = TRUE, names.eval = NULL, ...)

network.adjacency(x, g, ignore.eval = TRUE, names.eval = NULL, ...)

network.edgelist(x, g, ignore.eval = TRUE, names.eval = NULL, ...)

network.incidence(x, g, ignore.eval = TRUE, names.eval = NULL, ...)

Arguments

x

a matrix containing edge information

g

an object of class network

ignore.eval

logical; ignore edge value information in x?

names.eval

a name for the edge attribute under which to store edge values, if any

...

possible additional arguments (such as edge.check)

Details

Each of the above functions takes a network and a matrix as input, and modifies the supplied network object by adding the appropriate edges. network.adjacency takes x to be an adjacency matrix; network.edgelist takes x to be an edgelist matrix; and network.incidence takes x to be an incidence matrix. network.bipartite takes x to be a two-mode adjacency matrix where rows and columns reflect each respective mode (conventionally, actors and events); If ignore.eval==FALSE, (non-zero) edge values are stored as edgewise attributes with name names.eval. The edge.check argument can be added via ... and will be passed to add.edges.

Edgelist matrices to be used with network.edgelist should have one row per edge, with the first two columns indicating the sender and receiver of each edge (respectively). Edge values may be provided in additional columns. The edge attributes will be created with names corresponding to the column names unless alternate names are provided via names.eval. The vertices specified in the first two columns, which can be characters, are added to the network in default sort order. The edges are added in the order specified by the edgelist matrix.

Incidence matrices should contain one row per vertex, with one column per edge. A non-zero entry in the matrix means that the edge with the id corresponding to the column index will have an incident vertex with an id corresponding to the row index. In the directed case, negative cell values are taken to indicate tail vertices, while positive values indicate head vertices.

Results similar to network.adjacency can also be obtained by means of extraction/replacement operators. See the associated man page for details.

Value

Invisibly, an object of class network; these functions modify their argument in place.

Author(s)

Carter T. Butts buttsc@uci.edu and David Hunter dhunter@stat.psu.edu

References

Butts, C. T. (2008). “network: a Package for Managing Relational Data in R.” Journal of Statistical Software, 24(2). https://www.jstatsoft.org/v24/i02/

See Also

Examples

#Create an arbitrary adjacency matrix
m<-matrix(rbinom(25,1,0.5),5,5)
diag(m)<-0

g<-network.initialize(5)    #Initialize the network
network.adjacency(m,g)      #Import the edge data

#Do the same thing, using replacement operators
g<-network.initialize(5)
g[,]<-m

# load edges from a data.frame via network.edgelist
edata <-data.frame(
  tails=c(1,2,3),
  heads=c(2,3,1),
  love=c('yes','no','maybe'),
  hate=c(3,-5,2),
  stringsAsFactors=FALSE
  )

g<-network.edgelist(edata,network.initialize(4),ignore.eval=FALSE)
as.sociomatrix(g,attrname='hate')
g%e%'love'

# load edges from an incidence matrix
inci<-matrix(c(1,1,0,0, 0,1,1,0, 1,0,1,0),ncol=3,byrow=FALSE)
inci
g<-network.incidence(inci,network.initialize(4,directed=FALSE))
as.matrix(g)

# load in biparite dataframe with weights
bipMat<-data.frame(
        event1=c(1,2,1,0),
        event2=c(0,0,3,0),
        event3=c(1,1,0,4),
        row.names=c("a","b","c","d"))
net<-network(bipMat,matrix.type='bipartite',ignore.eval=FALSE,names.eval='pies')
as.matrix(net,attername='pies')

network

Classes for Relational Data

v1.16.1
GPL (>= 2)
Authors
Carter T. Butts [aut, cre], David Hunter [ctb], Mark Handcock [ctb], Skye Bender-deMoll [ctb], Jeffrey Horner [ctb], Li Wang [ctb]
Initial release
2020-10-06

We don't support your browser anymore

Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.