gendatamds {lazy.mds}

R Documentation

Generation of Distance Matrices with Error

Description

This function calculates the weighted Euclidean distance as:

d_ijk = sqrt( sum_a w_ka (x_ia - x_ia)^2 )

Note that W is NOT squared above.

Usage

gendatamds(n, nG = 1, ndim = 2, stderror = 0.2, X = NULL, W = NULL, print = 0)

Arguments

n	# of stimuli
nG	# of subjects
ndim	# of dimensions
stderror	The standard deviation of error to be added.
X	n x ndim Stimulus Configuration Matrix or NULL.
W	nG x ndim Subject Weight Matrix or NULL
print	= 1 to print the result

Value

A list of
vO: n*(n-1)/2 x nG array consisting of O_k, k=1,2, ..., nG
vD: n*(n-1)/2 x nG array consisting of O_k, k=1,2, ..., nG
X: n x ndim configuration matrix
W: nG x ndim matrix of subject weights
stderror: standard deviation of the addtive error

Examples

set.seed(1701)
n <- 15; nG <- 5; ndim <- 2
resgen1 <- gendatamds( n, nG=3, ndim=2, stderror=0.2 )
O <- vechinv(resgen1$vO, nodiag=1, array=1)  # This is n x n x nG.

X <- matrix( c( 0,4, 0,0, 3,0, 3,4, 1.5,2 ), ,2, byrow=1 )
dimnames(X) <- list(paste("s",1:nrow(X),sep=""),paste("d",1:ncol(X),sep=""))
W <- matrix( c( 1,1, 2,1, 1,2 ),,2, byrow=1 )
dimnames(W) <- list(paste("w",1:nrow(W),sep=""),paste("d",1:ncol(W),sep=""))
resgen2 <- gendatamds( X=X, W=W^2, stderror=0.2 )
O2 <- vechinv(resgen2$vO, nodiag=1, array=1)  # This is n x n x nG.

---