R/model-based.r

In clusterfly: Explore clustering interactively using R and GGobi

#' Display model based clustering with mvn ellipses.
#' Displays the results of model based clustering with an ellipse drawn from
#' the multivariate normal model for each group.
#'
#' @param model output from me function
#' @param data input data frame to me
#' @keywords cluster dynamic
#' @export
#' @examples
#' if(require("mclust")) {
#' eei <- me(modelName = "EEI", data = iris[,-5], z = unmap(iris[,5]))
#' vvv <- me(modelName = "VVV", data = iris[,-5], z = unmap(iris[,5]))
#' vvi <- me(modelName = "VVI", data = iris[,-5], z = unmap(iris[,5]))
#' mefly(eei, iris[,-5])
#' mefly(vvi, iris[,-5])
#' mefly(vvv, iris[,-5])
#' }
#' @importFrom plyr rbind.fill
mefly <- function(model, data) {
  mean <- model$parameters$mean
  var <- model$parameters$variance$sigma

  ellipses <- do.call("rbind", lapply(1:ncol(mean), function(i) {
    data.frame(ellipse(mean = mean[,i], cov = var[,, i], df=10), cluster=i)
  }))
  colnames(ellipses) <- c(colnames(data), "cluster")
  ellipses$TYPE <- factor("ellipse")
  data$TYPE <- factor("data")

  all <- rbind.fill(ellipses, cbind(data, cluster=max.col(model$z)))

  g <- ggobi(all)
  glyph_type(g[1]) <- c(1,6)[all$TYPE]
  glyph_colour(g[1]) <- all$cluster
  invisible(g)
}