tests/clusteringParametertest.R

In Rmixmod: Classification with Mixture Modelling

library(Rmixmod)
createGaussianParameter <- function(nbCluster, pbDimension) {
  proportions <- rep(1.0 / nbCluster, nbCluster)
  means <- matrix(rep(0, nbCluster * pbDimension),
    nrow = nbCluster,
    ncol = pbDimension
  )
  for (k in 1:nbCluster) {
    for (p in 1:pbDimension) {
      means[k, p] <- k
    }
  }
  variances <- list(c(1:nbCluster))
  for (k in 1:nbCluster) {
    variances[[k]] <- diag(pbDimension)
  }
  par <- new("GaussianParameter",
    proportions = proportions,
    mean = means, variance = variances
  )
  return(par)
}

data(iris)
data <- iris[, 1:4]

param <- createGaussianParameter(3, 4)
strategy <- mixmodStrategy(initMethod = "parameter", parameter = param)
out <- mixmodCluster(data = data, nbCluster = 3, strategy = strategy)