wilcox_R: What the package does (short line)

outmgvad <-
function(m,center=NA,plotit=TRUE,op=1,
xlab="VAR 1",ylab="VAR 2",rate=.05,iter=100,ip=6,pr=TRUE){
#
# Adjusts the critical value, gval used by outmgv,
# so that the outside rate per observation, under normality
# is approximately equal to the value given by the argument
# rate, which defaults to .05.
# That is, expected proportion of points declared outliers under normality
# is intended to be rate=.05
#
# When dealing with p-variate data, p>9, this adjustment can be crucial
#
m=elimna(m)
n=nrow(m)
newgval=sqrt(qchisq(.975,ncol(m)))
z=array(rmul(n*iter*ncol(m)),c(iter,n,ncol(m)))
newq=0
gtry=NA
val=NULL
for(itry in 1:ip){
newq=newq+9/10^itry
gtry[itry]=newq
}
gtry=c(.95,.975,gtry[-1])
if(pr)print("Computing adjustment")
for(itry in 1:ip){
for(i in 1:iter){
temp=outmgv.v2(z[i,,],gval=gval,op=op)$out.id
val[i]=length(temp)
}
erate=mean(val)/n
if(erate<rate){
newgval=sqrt(qchisq(gtry[itry],ncol(m)))
break
}}
res=outmgv(m,gval=newgval,plotit=plotit,op=op, xlab = xlab, ylab = ylab)
#list(results=res,used.gval=newgval)
list(n=res$n,n.out=res$n.out,out.id=res$out.id,keep=res$keep,used.gval=newgval)
}