README.md

In chenqi57/MMDS: Modified Multidimensional Scaling

MMDS

Modified Multi-Dimensional Scaling Algorithm.

Description

This package is an implementation of Modified Multi-Dimensional Scaling Algorithm.

Installation

Install MMDS from GitHub:

install.packages("devtools")
library(devtools)
devtools::install_github("chenqi57/MMDS")
library(mmds)

or directly from the MMDS_1.0.2.tar.gz:

install.packages("MMDS_1.0.2.tar.gz", repos = NULL, type = "source")
library(mmds)

Function

Three functions are contained in this package, where two functions are the implementation of the MMDS algorithm.

• eigen_centered: Center the sample matrix X to Y and return Y's corresponding distance matrix.
• MMDS: Modified Multi-Dimensional Scaling Algorithm using eigen-decompision methods provided by R.
• MMDS.cpp: Modified Multi-Dimensional Scaling Algorithm using eigen-decompision methods provided by C++.

Documentation

Documentations of the above functions can be accesed by typing ? before each function's name at the R command. For instance, the user can read the function MMDS's argument, output and examples in detail by typing ?MMDS.

Example

Suppose you have a 1000×1500 sample matrix sample, where the sample consists of 1500 data points from five guassian distributions with same covariance matrix 0.45I(1000) but different means. The information about which group each data point belongs to is stored in Labels.

MMDS

library(mmds)

sample_MMDS = MMDS(X = t(sample), MM = 2, sigma = sqrt(0.45), centered = FALSE)
data = data.frame(sample_MMDS)
data$label = Labels

ggplot(data,aes(x = data[, 1], y = data[, 2], colour = Labels)) + 
geom_point(size = 1) + xlab("") + ylab("") + theme(legend.position = "none")

MMDS.cpp

sample_MMDS_cpp = MMDS.cpp(X = t(sample), MM = 2, sigma = sqrt(0.45), centered = FALSE)
data2 = data.frame(sample_MMDS_cpp)
data2$label = Labels

ggplot(data2,aes(x = data2[, 1], y = data2[, 2], colour = Labels)) +
geom_point(size = 1) + xlab("") + ylab("") + theme(legend.position = "none")

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