Learn R Programming

multicross (version 2.1.0)

A Graph-Based Test for Comparing Multivariate Distributions in the Multi Sample Framework

Description

We introduce a nonparametric, graphical test based on optimal matching for assessing whether multiple unknown multivariate probability distributions are equal. This method is consistent, and does not make any distributional assumptions on the data. Our procedure combines data that belong to different classes or groups to create a graph on the pooled data, and then utilizes the number of edges connecting data points from different classes to examine equality of distributions among the classes. The functions available through this package implement the work described here: .

Copy Link

Version

Install

install.packages('multicross')

Monthly Downloads

34

Version

2.1.0

License

GPL (>= 2)

Maintainer

Divyansh Agarwal

Last Published

May 25th, 2020

Functions in multicross (2.1.0)

mcm

Multisample generalization of Rosenbaum's crossmatch test
multigene

Given two input matrices with the same number of observations but differrent number of variables, this function returns the largest canonical correlation between variables of matrix 1 (X) and those of matrix 2 (Y).
select_class

When the MCM/MMCM tests reject the null, class selection can help determine which of the K classes are the likely contributors for rejection
mhcccreate

Creates the null covariance matrix for mmcm, corresponding to the scenario when all K distributions are the same
mhccexecutelong

Calculates the pairwise crosscounts for the K classes being examined
split_mat

Split a data frame or matrix into subsets based on a particular categorical variable
mmcm

Use the Mahalnobis-type multisample test based on optimal matching to compare K different multivariate distributions