R/tropical.HC.AGNES.R

In TML: Tropical Geometry Tools for Machine Learning

#' Agglomerative (AGNES) tropical hierarchical clustering
#'
#' This function performs agglomerative (AGNES) hierarchical clustering over the
#' space of ultrametrics defining the space of equidistant trees
#'
#' @param D matrix of points defining a tropical polytope.  Rows are the
#'   tropical points
#' @param method linkage method: mean, min, or max
#' @return list of distances in when merges occur; list of indices of points in
#'   each cluster
#' @author David Barnhill \email{david.barnhill@@nps.edu}
#' @references David Barnhill, Ruriko Yoshida (2023). Clustering Methods Over
#'   the Tropically Convex Sets.
#' @export
#' @examples
#' \donttest{P <-Sim_points
#' Tropical.HC.AGNES(P, method=mean)
#' }

Tropical.HC.AGNES <- function(D, method=mean){
  ### method is one of `mean`, `min`, `max`
  d <- dim(D)
  distance <- rep(0, d[1])
  index.list <- list()
  index <- list()
  for(i in 1:d[1])
    index[[i]] <- c(i)
  index.list[[1]] <- index
  for(i in 2:d[1]){
    D1 <- make.list.matrices(D, index)
    best.pair <- finding.pair(D1, method)
    index[[best.pair[[1]][1]]] <- cbind(index[[best.pair[[1]][1]]], index[[best.pair[[1]][2]]])
    index <- index[-best.pair[[1]][2]]
    index.list[[i]] <- index
    distance[i] <- best.pair[[2]]
  }
  return(list(distance, index.list))
}