In nsdumont/jDirichletMixtureModels: Dirichlet Mixture Models for Clustering using Julia backend

knitr::opts_chunk$set(echo = TRUE)

Setup

library(jDirichletMixtureModels)

dmm.setup()

Example 1: Using BaseModel

An exmaple of creating a BaseModel object to use a bulit-in conjugate model and getting the cluster information:

X=rnorm(100)

m=dmm.BaseModel("UnivariateNormalModel", c(0.,1.,1.,1.))

o=dmm.cluster(m,X)

state=o[[1]]
dmm.summarize(state$clusters)


Xdata=rbind(matrix(rnorm(100),ncol=2),1+matrix(rnorm(100),ncol=2)/2)
m=dmm.BaseModel("MultivariateNormalModel", data=Xdata)

o=dmm.cluster(m,Xdata,iters=1000)

state=o[[1]]
dmm.summarize(state$clusters)

Example 2: Using JModel

An example of creating a JModel object: a model defined using user given Julia functions.

X=rnorm(100)
dmm.addfile("example2.jl")

m1=dmm.JConjugateModel("example_pdf", "example_post", "example_marg", list(0.0,1.0,2.0,0.5))
o=dmm.cluster(m1,X,iters=500)
dmm.summarize(o[[1]]$clusters)

m2=dmm.JNonConjugateModel("example_pdf", "example_pri", list(0.0,1.0,2.0,0.5))
o=dmm.cluster(m2,X,iters=500)
dmm.summarize(o[[1]]$clusters)

nsdumont/jDirichletMixtureModels documentation built on May 23, 2019, 2:51 p.m.