qgraph (version 1.6.2)

qgraph.layout.fruchtermanreingold: qgraph.layout.fruchtermanreingold

Description

This is a wrapper for the function that returns the x and y coordinates of the graph based on the Fruchterman Reingold algorithm (Fruchterman & Reingold, 1991), which was ported from the SNA package (Butts, 2010). This function is used in qgraph and is not designed to be used separately. See details for using constraints in this layout.

Usage

qgraph.layout.fruchtermanreingold(edgelist, weights=NULL, vcount=NULL,
	niter=NULL, max.delta=NULL, area=NULL, cool.exp=NULL, repulse.rad=NULL,
	init=NULL, groups=NULL, rotation=NULL, layout.control=0.5, constraints=NULL, 
	round = TRUE, digits = NULL)

Arguments

edgelist

A matrix with on each row the nodes at the start and the node at the end of each edge.

weights

A vector containing the edge weights.

vcount

The number of nodes.

niter

Number of iterations, default is 500.

max.delta

Maximum displacement, default is equal to the number of nodes.

area

The area of the plot, default is the square of the number of nodes.

cool.exp

Cooling exponent, default is 1.5.

repulse.rad

Repulse radius, defaults to the cube of the number of nodes.

init

Matrix with two columns and a row for each node containing the initial X and Y positions.

groups

See qgraph

rotation

See qgraph

layout.control

See qgraph

constraints

A constraints matrix with two columns and a row for each node containing a NA if the node is free or a fixed value for one of the coordinates.

round

Logical indicating if the initial input should be rounded

digits

Number of digits to round initial input and displacement in the algorithm to. Defaults to 5. This helps prevent floating point disrepancies between different operating systems.

Hard constraints

By using the 'constraints' argument the X and Y positions of each node can be fixed to a certain value. The 'constraint' argument must be given a matrix with two columns and a row for each node. An NA means that that coordinate for that node is free, and a value means it is fixed to that value.

Soft constraints

Soft constraining can be done by varying the 'max.delta' argument. This can be a single number, but also a vector containing the maximum displacement per step for each node. The default value is the number of nodes, so by setting this to a lower value for some nodes the node won't move so much. Use this in combination with the 'init' argument to make sure nodes don't move too much from their initial setup. This can be useful when adding a new node to an existing network and if you don't want the network to completely change.

Details

All arguments for this function can be passed from qgraph to this function by using the 'layout.par' argument, which must be a list containing the arguments. This can be used to constrain the layout in two ways:

References

Sacha Epskamp, Angelique O. J. Cramer, Lourens J. Waldorp, Verena D. Schmittmann, Denny Borsboom (2012). qgraph: Network Visualizations of Relationships in Psychometric Data. Journal of Statistical Software, 48(4), 1-18. URL http://www.jstatsoft.org/v48/i04/.

Carter T. Butts <[e-mail]> (2010). sna: Tools for Social Network Analysis. R package version 2.2-0. http://CRAN.R-project.org/package=sna

Fruchterman, T. & Reingold, E. (1991). Graph drawing by force-directed placement. Software - Pract. Exp. 21, 1129?1164.

See Also

qgraph

Examples

# NOT RUN {
# This example makes a multipage PDF that contains images
# Of a building network using soft constraints.

# Each step one node is added with one edge. The max.delta
# decreases the longer nodes are present in the network.

pdf("Soft Constraints.pdf",width=10,height=5)

adj=adjO=matrix(0,nrow=3,ncol=3)
adj[upper.tri(adj)]=1
Q=qgraph(adj,vsize=3,height=5,width=10,layout="spring",
	esize=1,filetype='',directed=T)
cons=Q$layout
for (i in 1:20)
{
	x=nrow(adj)
	adjN=matrix(0,nrow=x+1,ncol=x+1)
	adjN[1:x,1:x]=adj
	consN=matrix(NA,nrow=x+1,ncol=2)
	consN[1:x,]=cons[1:x,]
	layout.par=list(init=rbind(cons,c(0,0)),
	max.delta=10/(x+1):1,area=10^2,repulse.rad=10^3)
	y=sample(c(x,sample(1:(x),1)),1)
	adjN[y,x+1]=1
	Q=qgraph(adjN,Q,layout="spring",layout.par=layout.par)
	cons=Q$layout
	adj=adjN
} 
dev.off()
# }