MMDS: Modified Multi-dimensional Scaling Algorithm using R for...

In chenqi57/MMDS: Modified Multidimensional Scaling

Description

The function returns the MMDS embedding samples of X with embedding dimension MM.

Usage

MMDS(X, MM = 2, sigma, centered = TRUE)

Arguments

X	A d by n sample matrix with each column being a sample and each row being a variable.
MM	The embedding dimension, a positive integer from 1 to d. The default value is 2.
sigma	The noise standard deviation σ. If σ is unknown, the function will estimate it by eigen-values.
centered	Whether the sample matrix X is centered. The default value is TRUE.

Details

The sample matrix X is a d by n matrix, where d is the dimension of each observed data point and n is the number of such data points. The embedding dimension MM is a positive integer from 1 to d. The output will be a n by MM matrix which is the MMDS embedding samples of X with embedding dimension MM. The function itself will estimate sigma if the noise standard deviation σ is unknown, and will centeralize X if the user specifies X is not centered.

Value

A n by MM low-dimensional embedded sample matrix.

Examples

set.seed(496)
n = 10
d = 20
X = matrix(rnorm(d * n), n, d)
MMDS(X, MM = 2, centered = FALSE)

chenqi57/MMDS documentation built on Aug. 20, 2021, 6:34 p.m.