ldstatsHD (version 1.0.1)

plot.wfrl: Joint regression coefficient matrix estimator plot

Description

graphical representation of the non-zero joint regression coefficients structure

Usage

# S3 method for wfrl
plot(x, minn = 0, col = c("blue","red","green"), vertex.size = 2, 
      vertex.color = c("red", "blue"), edgesThickness = FALSE, 
      zoomThick = 10, ...)

Arguments

x

object of class wfrl.

minn

used visualization purposes in very dense networks. It only plots nodes that have degree larger than minn.

col

vector defining estimated edge colors: common edges (first element), only non-zero coefficients for first population (second element) and only non-zero coefficients for second population (third element).

vertex.size

plot.igraph parameter: vertex sizes.

vertex.color

vector defining the vertex colors for directed graph: first element describes the color of explanatory variables and second element describes the color for response variables.

edgesThickness

if TRUE, an edge thickness is proportional to the magnitude of its underlying estimated partial correlation coefficient.

zoomThick

it increases the thickness of all edges by zoomThick times (used for visualization purposes).

…

arguments passed to or from other methods to the low level.

Details

It produces a directed graph structure that connects explanatory variables to response variables.

See Also

wfrl for joint estimation of regression coefficients.

Examples

# NOT RUN {
N	<- 200
EX2 <- pcorSimulatorJoint(nobs = N, nclusters = 3, nnodesxcluster = c(60,40,50), 
                          pattern = "pow", diffType = "cluster", dataDepend = "diag", 
                          low.strength = 0.5, sup.strength = 0.9, pdiff = 0.5, nhubs = 5, 
                          degree.hubs = 20,  nOtherEdges = 30, alpha = 2.3, plus = 0, 
                          prob = 0.05, perturb.clust = 0.2, mu = 0, diagCCtype = "dicot", 
                          diagNZ.strength = 0.6, mixProb = 0.5, probSign = 0.7,  
                          exactZeroTh = 0.05)
					 
P           <- EX2$P
q           <- 50 
BETA1       <- array(0,dim=c(P,q))
diag(BETA1) <- rep(0.35,q)
BETA2       <- BETA1
diag(BETA2)[c(1:floor(q/2))]<-0
sigma2 	<- 1.3
Q       <- scale(EX2$D1)
W      	<- scale(EX2$D2)
X      	<- Q%*%BETA1 + mvrnorm(N,rep(0,q),diag(rep(sigma2,q)))
Y      	<- W%*%BETA2 + mvrnorm(N,rep(0,q),diag(rep(sigma2,q)))
D1     	<- list(scale(X),scale(Y))
D2     	<- list(scale(Q),scale(W))
## not run
#wfrl1   <- wfrl(D1, D2, lambda1 = 0.01, lambda2 = 0.05, automLambdas = TRUE, paired = FALSE, 
#               sigmaEstimate = "CRmad", maxiter=30, tol=1e-05)
#plot(wfrl1)
#plot(wfrl1, minn = 1, edgesThickness = TRUE)

# }