README.md

In yuanfang90/LNMVGA: Cluster microbiome / compositional data using a mixture of logistic-normal multinomial models

LNMVGA

This package provide a novel mixture of logistic-normal multinomial models for clustering microbiome data, or other compositional data. Parameter estimation are accomplished using variational Gaussian approximation to lift the computational overhead.

Methodology detail used in this package is from the manuscript in Fang and Subedi (2020). Datasets discussed in the simulation study of the manuscript is available in the data folder. Kxny.RData stores a collection of 100 replicated datasets of x-dimentional data with y observations.

Update 2023/12

Update to improve computation performance - Updated the following functions: + elbo_fun - matrix-wise computation replaced loop + LNM.clust - enabled parallelization

How to Install

The preferred way to install this package is through using devtools::install_github():

Preparations

If you haven't installed the package devtools, please do the following first:

install.packages("devtools")
library("devtools")

Install the package from GitHub

devtools::install_github("yuanfang90/LNMVGA", upgrade_dependencies = FALSE)

Then...enjoy using the package

library(LNMVGA)

yuanfang90/LNMVGA documentation built on Jan. 29, 2024, 8:24 a.m.