RMixtCompIO-package: RMixtCompIO

In RMixtCompIO: Minimal Interface of the C++ 'MixtComp' Library for Mixture Models with Heterogeneous and (Partially) Missing Data

RMixtCompIO

Description

MixtComp (Mixture Composer) is a model-based clustering package for mixed data originating from the Modal team (Inria Lille).

It has been engineered around the idea of easy and quick integration of all new univariate models, under the conditional independence assumption. Five basic models (Gaussian, Multinomial, Poisson, Weibull, NegativeBinomial) are implemented, as well as two advanced models (Func_CS and Rank_ISR). MixtComp has the ability to natively manage missing data (completely or by interval). MixtComp is used as an R package, but its internals are coded in C++ using state of the art libraries for faster computation. This package contains the minimal R interface of the C++ library.

Details

The main function is rmcMultiRun that runs a SEM algorithm to learn a mixture model.

See Also

rmcMultiRun. Other clustering packages: Rmixmod, blockcluster

Examples

dataLearn <- list(
  var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
  var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
)

dataPredict <- list(
  var1 = as.character(c(rnorm(10, -2, 0.8), rnorm(10, 2, 0.8))),
  var2 = as.character(c(rnorm(10, 2), rpois(10, 8)))
)

model <- list(
  var1 = list(type = "Gaussian", paramStr = ""),
  var2 = list(type = "Poisson", paramStr = "")
)

algo <- list(
  nClass = 2,
  nInd = 100,
  nbBurnInIter = 100,
  nbIter = 100,
  nbGibbsBurnInIter = 100,
  nbGibbsIter = 100,
  nInitPerClass = 3,
  nSemTry = 20,
  confidenceLevel = 0.95,
  ratioStableCriterion = 0.95,
  nStableCriterion = 10,
  mode = "learn"
)


# run RMixtComp in unsupervised clustering mode + data as matrix
resLearn <- rmcMultiRun(algo, dataLearn, model, nRun = 3)


# run RMixtComp in predict mode + data as list
algo$nInd <- 20
algo$mode <- "predict"
resPredict <- rmcMultiRun(algo, dataPredict, model, resLearn)

RMixtCompIO documentation built on Oct. 4, 2023, 1:07 a.m.