The mIDir package

An R package to implement the methodology described in A refreshing take on the inverted Dirichlet via a modeparameterization with some statistical illustrations (2025).

To install the package, use the following code in R

#install.packages("devtools")
library(devtools)
install_github("arnootto/mIDir")

Example

Code to reproduce Example 5.1: Cantaloupe spectra data (V3 vs V4)

# Fit the mIDir and cmIDir models to V3 and V4 of the cantaloupe spectra data
library(mIDir)
library(rrcov)
data("fruit")
X <- fruit[,c("V3","V4")]*10 #scale the data
a=ml.mIDir(X)
b=ml.cmIDir(X)

Code to reproduce Example 5.2: Australian Institute of Sport data

# Fit the mixtures of mIDir to the Australian Institute of Sport data
library(mIDir)
library(sn)
data("ais")
c=ml.MmIDir(X = as.data.frame(cbind(ais$LBM,ais$Wt,ais$BMI,ais$WCC,ais$Bfat)),k=2, method = "BFGS",initialization = "random.soft",max.iter = 1000)